Vol 16 No 1 - January-June 2016

International Journal of Computer Science
and Business Informatics

(IJCSBI.ORG)
ISSN: 1694-2507 (Print)

VOL 16, NO 1
ISSN: 1694-2108 (Online) JANUARY-JUNE 2016
IJCSBI.ORG
Table of Contents VOL 16, NO 1 JANUARY-JUNE 2016
Various Common Control Channel Assignment Schemes in CRNs: A Survey ......................................... 1

Jagsir Singh, Inderdeep Kaur Aulakh and Roopali Garg
Framework for Threat Modelling for a Power Utility: Case of Zimbabwe Power Utility Company .......... 8
Samuel Musungwini, Gilbert Mahlangu, Beauty Mugoniwa, and Samuel Simbarashe Furusa
Enhanced-bandwidth Compact Printed Inverted F Antenna Suitable for LTE/GSM Mobile Phone ....... 24
S. Jesse Jackson and T. Thomas
Implementation of e-Government Services using Java Technologies: A Methodological Guide in

Improving an e-Police Management System in Zimbabwe ...................................................................... 32
Mahlangu Gilbert, Chikonye Musafare, Furusa Samuel Simbarashe and Mugoniwa Beauty
Onto Tagger: Ontology Focused Image Tagging System Incorporating Semantic Deviation Computing
and Strategic Set Expansion ...................................................................................................................... 44
Gerard Deepak and Sheeba Priyadarshini J
International Journal of Computer Science and Business
Informatics
IJCSBI.ORG
Various Common Control Channel

Assignment Schemes in CRNs:
A Survey
Jagsir Singh
Research Scholar(ME-IT)
UIET, Panjab University Chandigarh, India
Inderdeep Kaur Aulakh

Assistant Professor & Coordinator (IT)
Roopali Garg
Assistant Professor (IT)
Abstract
Cognitive radio technology makes the unlicensed user to use the licensed band in
opportunistic manner. By using licensed band in intelligently way for solving the problem
of shortage of radio bands and optimal utilization of licensed band there is the requirement
control channel (CCC). Control channels are used for exchanging control information about
the new available channel for data transmission. When the primary user is detected or
decline of QoS of currently being used channel by cognitive user then there is need of the
common control channel to move on a new available channel. These channels are required
to sense the free channel, to transmit the control information for channel synchronization
between sender and receiver. It is challenging aspect of CRNs to provide the proper
coverage of CCC that the problems of data packet loss or termination of connection can be
reduced. In this paper various control channel assignment schemes are described.
Keywords
Common Control Channel (CCC), Media Access Control (MAC), Cognitive Radio
Networks (CRNs), Primary Users (PUs).
1. INTRODUCTION
Cognitive radio allows to cognitive users (unlicensed users) to use the
licensed band in opportunistic way for addressing the problem of radio
spectrum scarcity and utilization of licensed radio bands. It implements the
idea of spectrum sharing which permits the cognitive users to share free
licensed band intelligently without disturbing primary users
communication in the course of cognitive radio phenomena. In cognitive
ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 1

Informatics
IJCSBI.ORG
radio networks there must be channel synchronization between transmitter
and receiver upon which channel data will be transmitted to avoid the dada
packet loss.
A common control channel in CRNs does the various operations
such as make the connection between transmitter and receiver, detection of
other neighboring cognitive users [1], [2].We can define the control channel
in a number of ways such as a dedicated frequency band [3] or a time slot in
a time division multiple access systems. And a frequency hopping sequence
can be a common control channel. In an intelligent cognitive radio,
availability of the spectrum shows the both spatial and temporal variations
due to primary user activity. Thats why, there is no assurance that a
specified frequency band will be accessible for transferring the control
signals. We refer this is the problem of allotment of the common control
channels for controlling purpose.
In the cognitive radio networks common control channels
assignment is done in basic two ways static and dynamic channel allocation
algorithms. In this paper we will discuss the various control channel
assignment algorithms. There are numbers of challenges in the cognitive
radio networks to implement the common control channel algorithms. Some
of major issues are robustness to primary user activity, and security to
protect the control channel from jamming attack [10]. It is the major issue
of cognitive radio networks over which many researchers have been
working and they have purposed various common control assignment
methods which are described in remaining portion of this paper.
2. VARIOUS COMMON CONTROL CHANNEL ASSIGNMENT

SCHEMES
The categorization of common control channel design is the easy way to
know the control channel assignment schemes in CR networks which are
shown below:
Fig. 1 Categorization of CCC design [7].

Informatics
IJCSBI.ORG
2.1 Overlay
In overlay based CCC scheme, the CCC is permanently or for the short term
allocated of the spectrum not used by the primary user. When the allocated
common control channels are affected by primary users activity then
cognitive radio users has to leave the control channel and create the
connection over a new common control channel.
2.1.1 Out-of-Band
In this scheme, licensed or unlicensed band can be used like a common
control channel to establish connection between sender and receiver, node
co-ordination and synchronization. It helps to maintain the signaling
overhead which will increase the performance of the cognitive radio media
access protocol. Due to this, most of the cognitive radio network protocols
use unlicensed common control channel. If we use the unlicensed dedicated
common control channel it may be cause of interference with other
communications like networks like Bluetooth which will degrade the
performance of CR networks [11]. Also, transmit on single dedicated
common control channel will be subject to security attacks like control
channel jamming and Denial of service attacks (DoS) which can jeopardize
the entire cognitive radio network with single point of failure [11]. Thus,
designing common control channel through out-of-band technique has
number of issues with such as security attacks, and interference with other
technologies.
2.1.2 In-Band
In this common control design scheme, different intervals of time are used
to transmit the data and control messages [7]. Currently, sequential and
group based CCC design is proposed for node co-ordination and
synchronization. In sequential based network setup, CR nodes have to first
sense free channel list and scan each and every channel to find the common
channel which takes long time.
2.2 Underlay
In underlay scheme, same radio band can be used by both primary user and
cognitive radio users as a control channel. Control messages are transmitted
in low power by using short pulses by using spread spectrum phenomena,
which are transmitted over a high bandwidth. Because if transmit the
control information over large bandwidth then it does not disrupt the
communication of primary users. Even though, cognitive radio control
messages and primary user data can be transmitted concurrently in a
licensed spectrum.

Informatics
IJCSBI.ORG
3. MAIN COMMON CONTROL CHANNEL DESIGN SCHEMES
3.1 Dynamic CCC assignment using Swarm Intelligence

Swarm intelligence CCC scheme is purposed to dynamically find and
handle control channel in CRNs that does not require explicit messaging
and synchronization of clocks of both sender and receiver [4]. It is
circulated, in-band and flexible design which handle the problem of primary
user movement. Cognitive radio user chooses the best available free channel
as CCC when primary user return back over that channel [10]. In addition, if
numbers of control channel are less then it becomes easy to maintain.
Moreover its benefits, but there is a problem in swarm intelligence common
control channel scheme it requires high maintenance mainly in the presence
of fast changeable primary user activity.
3.2 Cluster based CCC assignment

In this scheme the common control channel in cognitive radio networks are
based on the time and the space-altering spectrum slots. A cluster-based
common assignment scheme has been purposed that defines we can assign
different control channel in different clusters in the cognitive radio networks
[3]. We make the cluster by dividing the CRNs and the common control
channel from the every cluster is chosen from the free channels. If the each
cluster has large common free channels then it is beneficial such as (i)
cognitive user can move to a new control channel if the current channel
occupied by a primary user, (ii) it becomes cooperative sensing easier[3].
And to handle the situation when spectrum opportunities are highly
heterogeneous a new algorithm designed which is known as Spectrum-
Opportunity Clustering (SOC). This technique provides the solution to PU
activity problem. The major disadvantage of this scheme is that, to handle
signal exchange between the clusters, it needs high control and it is very
difficult to synchronize the nodes in the clusters.
3.3 Hybrid CCC based Media Access Control Protocol

To improve the performance in the cognitive radio networks with respect to
the synchronization among the nodes of the cognitive radio networks and to
handle hidden terminal problem, a hybrid CCC based scheme has been
purposed [11]. It provides solution to avoid collisions due to hidden
terminal in multi channel network, reduce the time for set of connections of
network, robust to primary user activity and provides protection against
intruders. In hybrid common control based media access protocol provides
solution for above issues of CRNs to improve the performance of cognitive
radio-MAC protocol. It is combination of two spectrum band one is 902
MHz and other is 420 MHz which is a Television White space spectrum
band. Both bands are collectively used as a CCC for exchanging the control

Informatics
IJCSBI.ORG
signal between sender and receiver and for dynamic primary user activity
[11].
3.4 Synchronized MAC Protocol

It has been purposed to keep away from the need of the dedicated control
channel and also sort out the problem of hidden terminal in multi-channel.
In this scheme we divide the total time in to fixed-time intervals [12]. And it
is defined at the initial state which time slot will be used by the cognitive
radio users for transmitting the control information. A dedicated common
control channel has many disadvantage, one major drawback is the wastage
of radio spectrum. Secondly, a control channel can be overburden when
cognitive user increases [6]. Also, it becomes vulnerable to denial of service
attack. The synchronized media access protocol is a slotted protocol. It
collaborate the control channel contact together with standard vacant data
channel [8]. A time slot can works like a channel, over which control and
data can be transmitted. It also provides the solution to problem of
saturation of common control channels and jamming. The main drawback of
SYN-MAC scheme is when a primary user occupies the control slot then it
blocks the control channel operation.
3.5 Sequence based Rendezvous Common Control Channel Design

In sequence based CCC assignment scheme control channel are allocated
according to a sequence which can be a random sequence or a predefined
channel hopping sequence. These sequences are constructed by taking the
permutation of available control channel and choose best one from them. A
cognitive radio user can use the different sequence for different neighboring
user and it possible as well that neighboring users may be use different
hopping sequence. Through sequence-based rendezvous scheme, it is
possible to: (i) create an upper limit to the time to make contact , (ii)
establish a priority order for channels in which rendezvous occurs; (iii)
reduce the expected TTR as compared to random rendezvous. In this
scheme the sequences are pre-defined in order to visit the available control
channels that reduce the time required to make the contact with receiver in
minimum time. Basically the purpose of this method by using the particular
sequences which are constructed in such a manner to minimize expected
time-to rendezvous to find the best available control channel. If the numbers
of available channels are large, then it takes long time to find a adjoining
node on a channel for control information exchange [7].But there is a
problem in sequence based CCC assignment scheme that the order is
predefined and is not adjustable to new control channel opportunities.

Informatics
IJCSBI.ORG
3.6 Adaptive Multiple Rendezvous Control Channel assignment
It solves the problem of multiple rendezvous CCC scheme which is static in
nature. In this design the sequences can be chosen dynamically for handling
the issue of the interference with primary user. It achieves the high
performance by changing the hopping sequence when a primary is detected.
If we compare it with sequence based rendezvous scheme it does not require
strict synchronization. In simple terms it becomes easy to understand that, it
exploits the frequency hopping spread spectrum technique in which
transmitting channel changes over a particular time slice or according to a
pseudo random sequence. Similarly, this scheme is implemented in same
fashion to avoid the interference with other control channel and also provide
primary users activity robustness. But there is a problem in this scheme is
the average time to rendezvous may not be limited, cause of this time
required to create common control channel link may be long. This happens
when available channels are large [7].
4. CONCLUSION
To address the problem of common control channel assignment various
designs has been purposed by various researcher and computer scientists.
We have mentioned some of them. Each and every design algorithm has
own pros and cons. But the thing is that according to particular cognitive
radio networks we have to choose optimal common control channel
assignment scheme that it becomes easy to cognitive radio user to sense the
free available channel, exchange the control information related to channel
tuning and reduce the interference problem with other licensed(primary
user) and non-licensed(cognitive radio) users.
REFERENCES
[1] I.F. Akyildiz, W.-Y. Lee, M.C. Vuran, S. Mohanty, Next generation/dynamic
spectrum access/cognitive radio wireless networks: a survey, Elsevier Computer
Networks vol.50, no.13, pp. 21272159, Sept. 2006
[2] I.F. Akyildiz, W.-Y. Lee, K.R. Chowdhury, Cognitive Radio Ad Hoc Networks, Ad
Hoc Networks, Elsevier Ad Hoc networks, vol. 7, no.5, pp 810836, Jul. 2009.
[3] Sisi Liu, Loukas Lazos, and Marwan Krunz, Cluster-based Control Channel
Allocation in Opportunistic Cognitive Radio, Networks IEEE Transactions on
Mobile Computing, (TMC), Vol. 11, No. 10, pp. 1436 1449, 2012
[4] Christian Doerr, Dirk Grunwald and Douglas Sicker, Dynamic Control Channel
Assignment in Cognitive Radio Networks using Swarm Intelligence, IEEE Global
Communications Conference (Globecom) 2008.
[5] J. Zhao, H. Zheng, and G.-H. Yang, Spectrum sharing through distributed
coordination in dynamic spectrum access networks, Wireless Communications and
Mobile Computing Journal, vol. 7, no. 9, pp. 10611075, 2007.
[6] J. Zhao, H. Zheng, G.-H. Yang, Distributed coordination in dynamic spectrum
allocation networks, in: Proc. IEEE DySPAN 2005, pp. 259- 268, November 2005
[7] Brandon F. Lo, A survey of common control channel design in cognitive radio
networks, Elsevier Physical Comm. , vol.4, No-1, pp. 2639, March-2011

Informatics
IJCSBI.ORG
[8] C. M. Cordeiro and K. Challapali, C-MAC: A Cognitive MAC Protocol for Multi-
Channel Wireless Networks, in IEEE Symposium on New Frontiers in Dynamic
Spectrum Access Networks, April 2007.
[9] Yogesh R Kondareddy and Prathima Agrawal, Selective Broadcasting for Multi-Hop
CRNs, IEEE Sarnoff Symposium, Princeton, NJ 2008
[10] Claudia Cormio a, Kaushik R. Chowdhury, Common control channel design for
cognitive radio wireless ad hoc networks using adaptive frequency hopping,Elsevier
Ad-Hoc Nerworks Vol. 8, Issue 4, pp. 430-438, June 2010
[11] Satish Anamalamudi and Minglu Jin , Hybrid Common Control Channel Based MAC
Protocol for Cognitive RadioAd-Hoc Networks, International Journal of Information
and Electronics Engineering, Vol. 4, No. 3, May 2014
[12] Yogesh R Kondareddy and Prathima Agrawal, Synchronized MAC Protocol for
Multi-Hop CRNs, IEEE International Conference on Comm. Beijing, China 200.
This paper may be cited as:

Singh, J., Aulakh, I. K. and Garg, R. 2016. Various Common Control
Channel Assignment Schemes in CRNs: A Survey. International Journal of
Computer Science and Business Informatics, Vol. 16, No. 1, pp. 1-7.

International Journal of Computer Science and Business Informatics
IJCSBI.ORG
Framework for Threat Modelling

for a Power Utility: Case of
Zimbabwe Power Utility Company
Samuel Musungwini, Gilbert Mahlangu,
Beauty Mugoniwa, and Samuel Simbarashe Furusa
Computer Science and Information Systems
Faculty of Science and Technology
Midlands State University
Gweru Zimbabwe
ABSTRACT
The purpose of this study was analyse threats that are inherent in the prepaid electricity
meter system and to propose a framework for threat modelling. This framework can be
effectively used by power utilities power utilities in particular and other prepaid meter
system organisations to achieve end-to-end actionable insights on prepaid electricity
metering infrastructure. The study used a qualitative case research methodology with a
single unit of analysis. A purposive sampling technique was used to select suitable
participants. Data was collected from power utility engineers and security experts using
semi-structured interviews and focus group in order to triangulate the research findings.
The findings of the study indicated that at the present moment there are very few
frameworks that can be explicitly used to model threat to prepaid electricity infrastructure.
This has exposed the infrastructure to various attacks such as physical bypass, cyber-attack
and mechanically induced attack. We therefore recommend the adoption of an explicit
framework for modelling threat in prepaid metering infrastructure.
Keywords
Threat modelling, cyber attack, electricity theft, framework, ICTs, Mobile
technologies.
1. INTRODUCTION
The prepaid metering infrastructure is one of the most vital components of
the electricity grid system (Hmmerli, Svendsen & Lopez, 2013), because it
is the dominion of revenue collection for the power utilities. Its installation
within the electricity grid system has transformed electricity from the right
to use into a cash commodity which is one of its paramount characteristics.

IJCSBI.ORG
This has provided significant benefits to the power utilities, which include
maximising revenue collection, minimising cost associated with revenue
collection, reduced consumer debt on electricity usage, supplying purchased
electricity only, reduction in incorrect electricity billing, etc. (Miyogo,
Ondieki, & Nashappi, 2013). With the prepaid electricity metering
infrastructure, customers are now able to purchase electricity tokens from
various designated points such as power utilitys revenue halls and other
vending kiosks (Pabla, 2008). However, the installation of prepaid
electricity metering infrastructure to the consumer premises presents
different threats that have left it exposed/vulnerable to attack (Tondel,
Jaatun & Line, 2012).
An attack is an unwanted action that utilizes one or more vulnerabilities to
which when performed, it has the paossibility of compromising the purpose
of the installed infrastructure (Ucedavelez & Partner, 2012). In this case, the
purpose of the prepaid electricity metering infrastructure is to ensure that
consumers pay for electricity before use in order to reduce consumer debt,
among other reasons. The attack on prepaid metering infrastructure which
come in various forms and from different sources has seen power utilities
around the world losing millions of dollars through electricity theft. This has
negatively impacted revenue streams and backward operations of these
utilities.
2. BACKGROUD
In Zimbabwe, a prepaid electricity metering system was launched in 2012
for both domestic and commercial consumers to replace the conventional
post-paid meters and estimation billing (Megawatt Bulletin, 2012). The
concept of prepayment is built around paying before using a product or
service. Prepayment systems provide a disbursement for goods and services
before consumption or use (Casarin & Nicollier, 2010). In the context of
electricity distribution, the aspect of prepayment calls for the consumers to
hold electricity credit on their accounts (Miyogo, Ondieki & Nashappi.,
2013). The consumer can only use electricity as a commodity or service
when the account is paid up in advance.
The power utility has managed to install more than 900 000 prepaid
electricity meters for both domestic and commercial consumers by now
(Sibanda, 2014), with important applications already laid out and usable.
However, the installation of prepaid electricity metering system has seen the
power utility being deprived of millions of dollars by the consumers who
have found ways to steal from prepaid electricity meters. The power utility
is losing about US$10 million a month in revenue from electricity theft
(Share, 2014). Electricity theft has also seen the country experiencing more
loads shedding than ever in recent months, because the power utility cannot

IJCSBI.ORG
raise enough revenue for continuous electricity generations. The system was
introduced without proper feasibility study to check the polarity of the
prepaid electricity meters, hence, the consumers tampering with the system.
The power utility has reacted to electricity theft by introducing a Revenue
Protection Unit (RPU) to conduct field inspections/checks, investigate and
report consumers engaging in electricity thefts. Farawo & Towindo (2013)
reported that the power utility now offers monetary incentives as a way to
persuade consumers to provide information about electricity theft. Since it is
laborious to monitor all the 900 000 and more prepaid electricity meters to
be installed by means of field checks/inspection, there is need to protect the
system and ensure the security of the installed infrastructure (prepaid
electricity meter) so as to yield projected revenues for the investments made.
Although, electricity theft has been in existence since the 20th century, the
introduction of the prepaid metering system world over which have
transformed electricity into a cash commodity has seen perpetrators
adopting smart ways or methods of stealing electricity (Sreenivasan,
2011). Some of these methods are so complicated that they cannot be easily
detected by the current methods being employed by the power utilities. For
example the use of infra-red programming and micro-processor
programming techniques to alter electricity billing registers.
Therefore power utilities need to model threat to the prepaid electricity

metering infrastructure for them to figure out appropriate ways of combating
the attack. Threat modelling increases awareness of threat in order to
prepare for the security of the defined system. The threat should be
modelled so that all the possible attack strategies could be addressed using
appropriate mitigation methods. The process of identifying and discovering
vulnerabilities in an infrastructure requires awareness of the access points,
threats and their exploitations in achieving an attack goal (Zhang & Xu,
2006).
3. RESEARCH PURPOSE
The purpose of this study was to propose a unique and specific framework
for threat modelling that can be used by power utilities in an effective
manner to achieve end-to-end actionable insights on prepaid electricity
metering infrastructure. This study was guided by the following objectives:
Evaluate and analyse various models/frameworks used in threat
modelling.
Determine the steps and processes used in threat modelling.

IJCSBI.ORG
Establish vulnerabilities, threats and attack strategies in prepaid
metering infrastructure.
Identify sources of threats and types of attackers in prepaid metering
infrastructure.
Provide guidance for further inquiry ito threats prevention.
4. OVERVIEW OF THREAT MODELLING

Threat modelling is a method used by many security experts and
professionals to investigate security incidents and vulnerabilities in
information systems and infrastructure. According to (Microsoft, 2012), the
method involves identifying the infrastructure that needs protection from
attack and the potential threats to that infrastructure, classifying the threats
and finding mitigation strategies in a structured process. It is a practice
which serves as a organised approach in creating models that define, detect
and foretell threats to a given infrastructure in an organisation (Mcgrath &
Lennon, 2013). It encompasses the appreciative of the complexity of the
infrastructure and identifying all possible threats that may affect its purpose
of deployment and use (Zhang & Xu, 2006). Threat modelling can be
conducted before the deployment of the infrastructure as a preventive
measure or after the deployment as a way of maintenance.
The main factors covered in threat modelling should include identification
of what to be secured (system characterising), what an attacker may do
(threat), weaknesses (vulnerabilities) that gives the attacker intrusion
capabilities and the potential damages (Williams, 2007). The goal of threat
modelling is to identify and build possible threats and vulnerabilities to the
infrastructure in order to increase the security awareness of the organisation
and come up with adequate mitigation strategies. To further elaborate on
threat modelling, Ciampa (2009) cites that the objective of threat modelling
is to understand the goal of the attacker, the attacker himself, types of
attacks and the possible strategies that may be used to attack the
infrastructure. The modeller should study components like who is the threat,
where and how is it likely to occur and the vulnerabilities of the
infrastructure. Therefore, instead of relying with continuous discovery
mechanism, organisations can use threat modelling to mitigate and defeat
the attack scenarios before they fully develop (Hardy, 2012).
5. PRESENT APPROACHES TO THREAT MODELLING

The capability of an organization to respond to the emerging threats to their
infrastructure can be seen as a journey rather than a destination (Excellence,
2013), since attackers always aim to be ahead of the frameworks and models

IJCSBI.ORG
currently in use. Various frameworks and models have been used to model
threat to the infrastructure by different organisations. Some are simple and
ease to use by non-experts in the field of security while others are
complicated and require special skills and knowledge. These include, among
many the Threat Logic Tree, Attack Tree-Model, Game Theoretic Model,
PASTA model and Operationally Critical Threat, Asset, and Vulnerability
Evaluation (OCTAVE) framework. This study only looked at the Attack
Tree-Model, OCTAVE and Microsoft Security Development Lifecycle
(SDL) Threat Modelling. The decision to choose these models is because
ATM and Microsoft Security Development Lifecycle (SDL) Threat
Modelling have been used before to model threat in Advanced Metering
Infrastructure (AMI) and their concepts complement each other in threat
identification and evaluation. Although OCTAVE is yet to be used in AMI
and other metering infrastructure in electricity grids, its consideration was
based on the fact that it focuses on balancing operational risk, security
practices, and technology (Zhang & Xu, 2006), which is central to this
study.
5.1 Attack Tree-Model

This one of the first modelling techniques developed by Schneier (1999).
The model takes the form of a tree in demonstrating the attack whereby the
attack goal is the root node while the leaf nodes represents the steps to be
taken to accomplish the attack goal. It is a schematic mode of depicting how
attacks can occur to the infrastructure whereby the root node represent the
ultimate goal of the attack while leaf nodes represent the various ways an
attacker can use to achieve the ultimate goal (Ucedavelez & Partner, 2012).
The leaf nodes are decomposed until they reach a state where further
decomposition is no longer possible. Leaf nodes also represent the
condition(s) which must be fulfilled to accomplish successive goals or the
ultimate goal of the attack. This technique has been used to define attacks
against various information infrastructures and other real-world
applications. The technique has gained its popularity in computer science
and information systems research. It has been widely used by a number of
organisations to model threats to information technology and non-
information infrastructure. However, this model focused on either
classifying threats or modelling the behaviour of the attacker using one
specific factor which is the intrusion scenario. This has resulted in the
limited scope of the attack discoveries and mitigation strategies. There is a
need to extend the concepts covered in previous models to address factors
like threat rating, priorities, countermeasures and protection strategies.

IJCSBI.ORG
5.2 OCTAVE framework
OCTAVE is an approach that aims to increase the decision making process
of protection and management of organizational resources (Marek &
Paulina, 2006). It is intended to allow people to comprehend the security
matters of the organizations and work towards improving them with least
exterior assistance. It is underpinned on the philosophy and the principle of
self-direct which states that people from the organization are in a better
position to decide on the security of their resources. Thus, it provides a
systematic and context-driven approach for the organization to manage
threats to infrastructure. This technique studies, both technical and
organisational issues in order to portray the security needs of an
organisation using a three phase approach (Bakari, 2007) as indicated in
Figure 2.
Figure 1: The OCTAVE framework

Source: Marek & Paulina, 2006
As shown in Figure 2, the major components of the OCTAVE framework
are Building of Asset-Based Threat Profiles, Identifying Infrastructure
Vulnerabilities and Developing Security Strategy and Plans. Within each
phase, various processes and activities exist. The other factors that are
covered in these three phases though not visible include information
gathering, characterising the assets/infrastructure, describing threats and

IJCSBI.ORG
current mitigation strategies/plans and establishing security requirements.
The OCTAVE framework has presented the foundation and guidelines for
many organizations to deal with various threats. Although the OCTAVE
framework covers major factors for threat modelling, it focuses more on
information infrastructure. Therefore, there is a need for an improved
framework that will encompass the physical infrastructure that is deployed
external to the organization such as the prepaid electricity meter.
5.3 Microsoft Security Development Lifecycle (SDL) model

This model has been used in conjunction with the ATM in modelling Threat
in Advanced AMI by Tondel et al. (2012) in designing Demo Steinkjer
conceptual framework. It is based on two categories which are Threat
overview and Attacker strategies.
5.3.1 Threat overview

In classifying threat to the AMI, Tondel et al. (2012) used a STRIDE model
which is a borrowed concept from Microsoft to aid the modeller in
categorizing the threat. The model is concerned with determining the access
an attacker may have to the data (spoofing) as the meter communicates with
the back-end system, access to the configuration settings of the meter
(information disclosure) such as billing registers and communication link in
order to tamper with the billing cycle and the elevation of privileges
remotely or physically.
5.3.2 Attack strategies in AMI

These can be identified through information gathering from various sources
and presented together with their goals and the ways to achieve them. In
order to carry out an attack, the attackers need to have knowledge of the
power utilitys communication grid topology and meter configurations. In
addition, access is also necessary to plan the attack. Some of the attack
strategies identified by Tondel et al. (2012) include:
Manipulating power measurement (physically)
Manipulating measurement values
Manipulating messages from the meter
Physical break-in
Break-in via infra-red port
6. PROCESSES/STEPS IN THREAT MODELING

IJCSBI.ORG
The processes and steps to be followed in the threat model differ according
to the framework/model used. Although frameworks and models use
different processes/steps, they all intend to achieve a common goal that is to
provide guidelines in dealing with threats and attacks to the infrastructure
(Bertino, Martino, Paci, & Squicciarini, 2010). Figure 3 shows a six (6)
step-process model proposed by Meier et al. (2003) in their study on
Improving Web Application Security.
Figure 2: Process in threat modelling

Source: Meier et al. (2003)
6.1 Identify Assets

The major reason why threat and attacks exist is the availability of
assets/infrastructure in the organization. Therefore the goal of the attacker is
to gain access to a particular asset/infrastructure of interest and perform an
attack. Asset/infrastructure identification involves the classification of
critical/vital infrastructure or systems that an organization has to
protect/guard against attack. Sub-processes may include describing why the
asset/infrastructure needs to be protected and determining access points
available for the attack to be possible (Burns, 2005).
6.2 Create an architecture overview

The goal here is to identify the key functionalities, the characteristics and
the people served by the asset/infrastructure. It can be represented by Data
Flow Diagrams (DFDs), tables and sequence diagrams to document the
system and its subsystems. End-to-end deployment scenarios should also be
specified so that threat identification in step 4 becomes easy. Therefore, the

IJCSBI.ORG
more the modeller has knowledge about the key functionalities,
characteristics and the people served by the asset/system, the easier it is to
find threats and determine vulnerabilities
6.2.1 Decompose the application
In order to focus the attention of threat modelling to the areas of concern,
the architecture should be divided into sub components. This is the third
step that provides an illustration of the boundaries of the asset to help define
the parameters of threat modelling.
6.2.2 Identify the threats.

A threat is a potential occurrence that can compromise the
asset/infrastructure. For it to occur, they should be a target
asset/infrastructure and a vulnerable point. In performing threat modelling,
the modeller should use a systematic approach to discover all vulnerabilities
of the asset/infrastructure that an attacker can be exploited to achieve the
goal. They should be identified in view of the goals of an attacker,
knowledge of the asset and potential vulnerabilities to the asset. It can be
identified by using the attack trees, STRIDE model during a brainstorming
session or just using the knowledge of the asset to list the ways in which the
attacker can achieve the goals.
6.2.3 Document the threats.

A template can be created and used to document each based on certain
attributes. Some of the attributes include threat description and threat target.
Other attributes may include the attack techniques which may show the
vulnerabilities exploited. This is the final stage of threat modelling.
6.2.4 Rate the threats.

The threat should be rated by assigning the probabilities of damage should
they occur. This enables the organization to give priority to the most risk
threats since it is not possible and economically viable to address all the
threats. The threat should be rated in accordance to the risk they pose to the
asset/infrastructure and its severity. It can be rated using the damage
potential, reproducibility, exploitability, affected users and visibility.
7. METHODOLOGY
A qualitative case research methodology (Yin, 2014), with a single unit of
analysis was used in this study to collect data from various sources. The

IJCSBI.ORG
unity of analysis in this study was a power utility which has installed
prepaid metering infrastructure. A purposive sampling technique was used
to choose participants from the target population. The sample of this study
was comprised of six (6) electrical engineers who have worked in the power
utility for more than three (3) years. Data was collected from both the power
utility engineers and security experts using semi-structured interviews and
focus group, respectively in order to triangulate the research findings and
evaluate the framework
8. FINDINGS AND DISCUSSION

8.1 Comparative analysis of available frameworks and models
This study commenced with an extensive literature review to establish the
current frameworks and models used in threat modelling by various
organisations. Although a number of them exist, the study found that most
of the frameworks and models are directed at solving threats on IT and
physical infrastructure that is located within the organisation. The
frameworks and models reviewed in the literature and presented in Section 2
have limited capabilities in dealing with the threat to organisational
infrastructure that is deployed in the front-end like the prepaid electricity
meter. Apart from focusing on addressing threats to internal systems and
infrastructures, it can be argued that the Attack-Tree Model and the
Microsoft Security Development Lifecycle (SDL) model are not suitable
enough for threat modelling. These approaches only assist the organization
to identify ladders and routes an attacker can use to reach the goal.
8.2 Vulnerabilities, threats and attack strategies

The findings of the study have presented the landscape for the authors to
understand vulnerabilities, threats and attack strategies to the prepaid
electricity metering infrastructure. The prepaid metering infrastructure is
highly vulnerable to attack because the attacker has unlimited physical
access since they are located in consumer premises. Therefore, when an
organisation is carrying out threat modelling to infrastructure it should
consider the sources of threats, types of attackers, the level of access they
have and the skills of the attacker. According to the findings of the study,
prepaid metering infrastructure is vulnerable to three types of threats. These
are:
Threat to wiring integrity- this can be exploited through partial and
wholly bypass whereby the cable can be either diverted from the
service line to the load or the prepaid electricity meter is totally
disconnected from the electricity grid system. Here the attacker uses
a physical bypassing strategy. This is the easiest strategy that an

IJCSBI.ORG
attacker can be used because no skill is required to exploit the threat.
The attacker only needs to be brave and an understanding of the
wiring integrity of the prepaid electricity meter. The attacker may
use two approaches which are visible and non-visible approach. In
the former the bypassing is done outside the prepaid electricity meter
while in the latter bypassing takes place inside the meter. The goal is
to divert the flow of electricity from the service line to the load so
that electricity consumption is not measured.
Threat to billing software embedded in the prepaid electricity meter
(Cyber-attack strategy)- this threat can be exploited by using an
infrared device or micro-controller programming which access the
billing software through the port of the meter. Since every prepaid
metering infrastructure has an infra-red port to allow the electrical
engineers to carry out maintenance work, the ports permit access to
any device that uses the infra-red light. The attack is non-visible and
non-physical because it cannot be detected by just a mere
observation and the attacker does not need to have a physical contact
with the infrastructure. This strategy is usually used by intelligent
people and other experts in computer programming. The goal of this
strategy is to reduce the billing cycle so that few units of
consumption will be recorded by the infrastructure.
Threat to the mechanically built-up (Mechanically induced strategy)
- particularly the meter disk, the magnet and other internal
components. The meter disk is the most important component of the
infrastructure that enables power utility record the electricity that has
been consumed. This threat can be exploited by inserting metal
objects to the infrastructure or exposing it to a strong magnetic
object so that the disks rotary movement is not proportional to the
electricity consumed. The attack can be visible or non-visible. Any
potential attacker can use this strategy since its not complicated.
The goal of this strategy is to make the disk to rotate slowly so that
less electricity is recorded. The slow the movement of the disk the
low electricity is recorded.
An attack-tree model shown in Figure 3 below has been used to model the
vulnerabilities, threats and attack strategies to prepaid electricity metering
infrastructure.
As shown in Figure 3 below, the tree starts by defining the goal of the
attacker which is the root node. In this case the ultimate goal of the attacker
is to defeat the electricity billing or enjoy free electricity. The nodes that
follow depict the strategies that can be used to fulfil the ultimate goal. The
strategies are decomposed further down so that all the possible

IJCSBI.ORG
actions/threats that may lead to the accomplishing of the goal are known.
The decomposition process will continue until the vulnerabilities are
identified which make it impossible to break the tree any further.
Defeat electricity billing/enjoy free electricity consumption
Physical bypass Cyber-attack Mechanical induced
Partial bypass Wholly bypass Microcontroller prog. Infra-red prog. Mechanically

interference
Crossover Total meter Cable Neutral Phase Magnetic Strong heat Objects
cabling disconnecti alternation removal diversion influence exposure injection
ng
Figure 3: Outline of intrusion scenarios for prepaid electricity metering

infrastructure
Source: Researchers own construction
9. SOURCES OF THREATS AND TYPES OF ATTACKERS
Threat sources can either be internal or external
Power utility employees (current and former)
Technologically gifted people/experts
Experimental people
Financially disadvantage people
10. PROPOSED FRAMEWORK

The proposed framework presented in Figure 4 has been conceptualized to
consist of three (3) phases:
Phase 1: Attack detection/discovery- this is the initial stage of
modelling a threat. The modeller should check for any information
that may give clues to the attack, i.e. there is an attack that is going
on. In the case of the prepaid electricity meter infrastructure,
electricity purchasing reports can act as a lead, especially if the
purchasing trends have dropped to low levels. Comparison of
consumer and purchasing reports may also be used to guide the
modeller for the perceived attack on the infrastructure. With lead

IJCSBI.ORG
information, the modeller can consult employees and security expert
on more clues. After that a security survey/field inspection can be
conducted to confirm or refute the attack by looking for visible and
invisible attacks. Lastly, there is a need to recognise and generate the
attack patterns based on the methods used.
Phase 1 Check for lead information: Brainstorm (internal/external):

Review performance report Employees
List potential attacks Security experts
Attack
Conduct security survey/field
detection/discovery Recognize & generate attack patterns: inspections:
Method commonly used Check for visible & non-
No. of occurrences visible attacks
Outline intrusion/attack scenarios: Describe attack goal (s):

Technological Immediate goals
Phase2 Non- technological Ultimate goal
Map intrusion & attack scenarios to attack goals using attack tree (s)
Define & profile potential attackers: Rate threats: (PIVE)

Threat assessment, Skills Possibility
analysis & evaluation Characteristics Impact (organization & customer)
Capabilities Visibility
Motivation Exploitation
Mitigation: Define protection strategy:

Figure Examine current strategies
4:3Proposed framework
Phase Technological
List protection gap Non-technological
Prioritize threat
Check for infrastructure
dependencies
Develop mitigation
plan & strategy
Resolve vulnerabilities & review progress
Figure 4: Proposed framework

IJCSBI.ORG
Source: Own construction
Phase 2: Threat assessment, analysis & evaluation- this is the main
phase of threat modelling. It requires an outline of the
intrusion/attack scenarios based on technological and non
technological conduct. For example, cyber-attack is technological
while the physical by-pass and mechanically induced are both non-
technological. The phase also requires the description and mapping
of attack goals to intrusion scenarios, profiling attackers in order to
understand their capabilities. Finally, the threats need to be rated
using the PIVE model which has been proposed by this study.
Phase 3: Develop mitigation plan & strategy- this phase involves
the examination of current protection strategies in order to identify
gaps, prioritise the threat and checking for infrastructure
dependencies. The modeller also needs to decide on the protection
strategy based on the factors like cost befit analysis in order to
decide whether to respond with technology or not. Last
vulnerabilities should be resolved and a review carried out.
11. LIMITATIONS
This research was conducted at a time when the Prepaid metering system
has just been rolled in Zimbabwe and as such this being a new system in a
new environment, some threats are still to be discovered. Hence these
researchers believed this framework to be a good starting point. In this
manuscript the authors have confined the framework to ZESA the power
utility organisation in Zimbabwe but in reality we believe this framework
can also be extended to other organisations in Zimbabwe and beyond.
12. CONCLUSION
Threat modelling is one of the key requirements that enable organizations to
minimize security risks and achieve operational excellence. When
conducting threat modelling, the modeller should view the infrastructure
from the position of the attacker. This will enable them to see it as an
exposed system. Power utilities can use threat modelling to mitigate and
defeat attack scenarios before they can cause severe damage.
13. FUTURE RESEARCH

Having proposed the framework for threat modelling to the prepaid
metering infrastructure, the guidance for further inquiry may commence by
looking at the strategies used to steal electricity in order to reinforce the
proposed framework. The research may also proceed by establishing the

IJCSBI.ORG
strategies used by power utilities in dealing with the attack to prepaid
metering infrastructure. Furthermore, the use of mobile technologies to
achieve end-to-end actionable insights and deliver operational analytics on
the infrastructure may also be considered.
14. ACKNOWLEDGEMENTS
The researchers would like to acknowledge the valuable contributions to the
success of this work by people from the power utility in Zimbabwe and
registered electricity consumers and the framework reviewers. We would
like to recognize the assistance rendered by these as their valuable input
enabled us to craft this piece of work. We are indeed grateful to these
participants for availing themselves for the research and providing data that
enabled this study to be carried out. Otherwise this research would not have
been possible. Their priceless contribution is greatly appreciated.
REFERENCES
Bakari, J. K. (2007). A Holistic Approach for Managing ICT Security in Non-Commercial
Organisations: A Case Study in a Developing Country.
Bertino, E., Martino, L. D., Paci, F., & Squicciarini, A. C. (2010). Security for web services
and service-oriented architectures. Security for Web Services and Service-Oriented
Architectures, 1226. http://doi.org/10.1007/978-3-540-87742-4
http://www.arrow.dit.ie/cgi/viewcontent.cgi?article=1012&context=engschcivcon
Burns, S. F. (2005). Threat Modeling: A Process To Ensure Application Security, (January)
SANS Institute
Ciampa, M., (2009). Security and Guide to Network Security Fundamentals. 3rd edition.
Boston: Cengage Learning.
Excellence, N. (2013). Deliverable D6 . 3: Advanced Report on Smart Environments.
Hmmerli, B., Svendsen, N. K. & Lopez, J. 2013. Critical Information Infrastructures
Security: 7th International Workshop, CRITIS 2012, Lillehammer, Norway, September 17-
18, 2012. Revised Selected Papers, Springer Berlin Heidelberg.
Hardy, G. M. (2012) Beyond Continuous Monitoring: Threat Modeling for Real-time
Response, (October) SANS Institute
Marek, P., & Paulina, J. (2006). The OCTAVE methodology as a risk analysis tool for
business resources. International Multiconference Computer Science and IT , 485497.
Retrieved
from:http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:The+OCTAVE+me
thodology+as+a+risk+analysis+tool+for+business+resources#2
Mcgrath, M., & Lennon, R. (2013). Letterkenny Institute of Technology Threat Modelling
for Legacy Enterprise Applications, (August).
Meier, J. D., Mackman, A., Dunner, M., Vasireddy, S., Escamilla, R., & Murukan, A.
(2003). Chapter 3 Threat Modeling, (June).

IJCSBI.ORG
Microsoft. (2012). Introduction to Microsoft Security Development Lifecycle (SDL) Threat
Modeling. Retrieved from http://www.cs.berkeley.edu/~daw/teaching/cs261-
f12/hws/Introduction_to_Threat_Modeling.pdf
Miyogo, C. N., Ondieki, S., & Nashappi, G. (2013). An Assessment of the Effect of
Prepaid Service Transition in Electricity Bill Payment on KP Customers , a Survey of
Kenya Power , West Kenya Kisumu, 3(9), 8897.
Pabla A.S., (2008). Electric Power Distribution, 5th edition. New Delhi: Tata McGraw-Hill.
Available at:
http//www.books.google.com/books/about/Electric_Power_Distribution.html?id...
Schneier, B. (1999). Attack Trees. Dr Dobbs Journal, 24(12), 2129. Retrieved from
http://www.schneier.com/paper-attacktrees-ddj-ft.html
Tndel, A. I. , Jaatun, M. G. & Line, M. B. (2012). SecurityThreats in
DemoSteinkjer_v1.pdf. SINTEF ICT and Telenor, 1.
Ucedavelez, T., & Partner, M. (2012). Real world threat modelling using the pasta
methodology. pp.2-61. Available at:
https://www.owasp.org/images/a/aa/AppSecEU2012_PASTA.pdf
Williams, L. (2007). Threat Models Software Security Touchpoints: Purpose of Threat
Modeling, 115.
Yin, R.K. (2014) Application of case study research: Design and Methods. 5 th edition.
London: Sage publications.
Zhang, X., & Xu, S. (2006). TDDC03 Projects , Spring (2006): A Comparison of Attack
Trees , Threat Modeling and OCTAVE.

Musungwini, S.,Mahlangu, G.,Mugoniwa, B., and Furusa, S. S., 2016.
Framework for Threat Modelling for a Power Utility: Case of Zimbabwe
Power Utility Company. International Journal of Computer Science and
Business Informatics, Vol. 16, No. 1, pp. 8-23.

IJCSBI.ORG
Enhanced-bandwidth Compact
Printed Inverted F Antenna
Suitable for LTE/GSM Mobile
Phone
S. Jesse Jackson
PG Student, ECE Department
QIS College of Engineering and Technology, Ongole, India
sikha.jessejackson@gmail.com
T. Thomas
Assistant Professor, ECE Department
QIS College of Engineering and Technology, Ongole, India
thomas.thatapudi.t@ieee.org
ABSTRACT
In this paper design of multi band Printed Inverted F Antenna for LTE/GSM mobile
application is manifested. The proposed design is suitable for embedding in the mobile
handset and has advantages in aspects like low profile, small radiation effect, and good
tuning properties. A matching circuit is employed in addition with the proposed aerial
constructed by a rectangular plate placed on top of the FR-4 dielectric substrate of 65x80
mm2 size to provide wide operating frequency band. This makes the design more suitable
for better tuning to a required frequency bands. The ground plate is coated on substrate of
its bottom side. Overall size of PIFA is small enough to be embedded in a smart phone, and
it has been optimized to cover the LTE and GSM bands at 2600, 1500, and 800 MHz
frequencies respectively. It has a gain ranging from 1.469 dBi to 5.428 dBi at the resonating
frequency bands. The study of antennas performance in term of impedance bandwidth, S-
parameters, radiation pattern and gain is presented with CST Microwave studio simulation
results.
Keywords
Chip inductor, chip capacitor, Impedance matching circuit, Mobile phone, Multi band,
PIFA, Tuning.
1. INTRODUCTION
Nowadays mobile phones are everywhere and users are depending on
phones for multiple services. The hard-hitting issues for handheld cellular
user equipment are the antennas with broad impedance bandwidth, low
profile, less weight and it should be cost effective, and Omni-directional
radiation pattern is desirable. To provide continuous interactive voice, data

IJCSBI.ORG
and even video services many improvements are being made in
communication systems to make these services available anytime and
anyplace. Next-generation cellular communication systems higher
bandwidths are mandatory, LTE bands are only alternative for future needs.
LTE generally offer high reception date rates [13]. In general based on
specification either internal or external antennas are used in mobile
handsets. The close proximity of external antenna causes higher radiation
absorption rate, it is the major disadvantages of such type of mobile
antennas. Contrary to this, to avoid human interference internal antennas are
printed on substrate system board, and care will be taken to make sure that
the major radiation from antenna is radiated away from head. The printed
aerial are designed depending on the mobile phone type and amendments
have to be made to meet the specifications. The design of low profile
antenna is expected in the mobile handsets and all performance parameters
should be up to the mark so that it can be used for commercial services. To
achieve wide bandwidths for LTE/ WWAN operation the inverted-F
antenna (IFA) is usually used as internal antenna in mobile phones, but 6-
10mm of extra profile is a added size to the basic design, leading to more
sophisticated design called PIFA printed inverted F-shape antenna [4-9].
Mono pole antenna is improved in many aspects to be used in mobile
phones, which is now called as PIFA. The advantages of PIFAs made them
very suitable for small devices so they are widely used in mobile user
equipment. A wide planar radiator is used in place of a slim conductive
radiating element of an Inverted F antenna, this new design is named as
Planer IFA. In the basic PIFA construction one element as top plate, another
as ground plane, for feeding the resonating upper element a strip is used. For
DC-shorting between ground and upper element a plane is used for
connection at one edge of the radiating plane. For better impedance
matching, the following are optimized in PIFA design, they are the signal
feed position, and the shorting pin position in the slot. The separation
between signal feed and DC-shorting pins is another design issues for the
better impedance matching of the PIFA. Design of both E-plane and H-
plane polarizations with desired directional properties is possible with PIFA.
PIFAs are the best antennas when the antenna point of reference is not
predictable and reflections are present [10-12]. Control of the resonance
frequencies independently by changing lengths of strips or slots is a widely
known method in the micro strip patch antenna design.
The antenna space in a mobile handset environment is extremely limited.
Therefore, designing an internal antenna for a mobile handset is difficult,
especially when multi-band operation is required. This paper presents an
antenna for mobile phone with attractive characteristics as shown in figures
1 and 2, the proposed antenna [13] with multiband operation design to

IJCSBI.ORG
support GSM/LTE bands is presented. Embedded matching circuit provides
tuning property for the system to different resonant frequencies.
Figure 1. Proposed antennas radiating section with slots and matching circuit
2. PROPOSED ANTENNA
Tuning of antenna for different resonance frequencies is easy with alteration
of element values of a matching circuit connected to the antenna of a mobile
phone. To avail this advantage in the proposed design matching circuit with
three elements is employed. A capacitor (1 pF) is connected in series with a
parallel combination of capacitor (3.1 pF) and inductor (8.8 nH). For the
given set of values of matching elements the proposed antenna resonates at
three different frequencies. And other lower and upper resonance
frequencies can be obtained with different combination of values for the
matching circuit elements.
The substrate is a made up of FR4 having r = 4.4 and thickness 0.8 mm, on
top of this the radiator plane of 35 x 10mm2 surface area is realized. At a
height of 8.2 mm the separate PCB structure is hovered above a ground
plate of 65 x 80 mm2 with the support of DC-shorting rod with 1.2mm
radius and 9mm height. Slots in the structure are crafted to make it resonate
at different frequencies, moreover the spacing between feed and shorting
strips has considerable impact on resonance frequencies.
For resonance at different frequencies patch slots are introduced. In figure 2.
(b) the measurements of the slots of width 1mm in the radiator patch are
shown. For the feed line the length and the width are 15.5mm and 1mm.
Where as in case of rectangular patch antenna dimensions are calculated
based upon operating frequency. Dimensions of radiator are calculated from
the following equations [13], in which Fr resonant frequency, c light
velocity, permittivity of substrate. To make available higher data rates for
various services the antennas are planned to support foremost cellular
system bands namely GSM/UMTS/LTE. In this proposal both rectangular
strip and its slots are designed to support multiple bands.

IJCSBI.ORG
C
L (1)
2 Fr eff
1 2
W (2)
2 Fr o o r 1
(a)
(b)
(c)
Figure 2. (a) Proposed antenna with elevated radiator shorted to ground. (b)
Dimensions of radiating element (all dimensions are in mm). (c) Matching circuit for
the proposed antenna, the rating of the capacitor is pF
while that of the inductor is in nH.

IJCSBI.ORG
3. SIMULATION RESULTS
When matching circuit with given combination of elements is connected
resonance can obtained at three different frequencies that is 0.8 GHz, 1.5
GHz and 2.7 GHz as shown in Fig. 3. On the other hand the antennas return
loss value with coaxial feed without matching circuit is -15.5dB at 0.93GHz
and -31dB at 1.8 GHz. The impedance bandwidth is wide enough to cover
GSM800, 1500 and LTE2600 bands. The proposed Z-shaped feed extension
helps to adjust the return loss value to a desired lower level and shows
considerable affect on bandwidth characteristics of the proposed antenna.
Figure 3. S11 return loss value over 0- 3GHz frequency band.
(a) (b) (c)

Figure 4. 3D Radiation pattern of the proposed antenna at
(a). 0.8GHz, (b). 1.5GHz, (c). 2.7GHz

IJCSBI.ORG
(a) (b) (c)

Figure 5. 2-D Radiation pattern at (a). 0.8GHz, (b). 1.5GHz, (c). 2.7GHz
The radiation pattern can be defined as the angular distribution of the
strength of the radiator. The 2D and 3D radiation patterns are shown in
figure 5 at the 0.8GHz, 1.5GHz and 2.7 GHz. From given figures it is
obvious that the proposed antenna exhibits omni-directional radiation
characteristics at lower resonant frequency, where as its radiation
characteristics are different from omni-directional at higher frequencies. For
the given set of frequencies the proposed antenna offers directive gain
ranging from 1.469 dBi to 5.428 dBi, the substrate height, thickness of the
substrate, and relative permittivity of the substrate can affect the gain
characteristics of the proposed antenna.
(a) (b) (c)

Figure 6. Current density at (a). 0.8GHz, (b). 1.5GHz, (c). 2.7GHz
The surface current distribution at different resonance frequencies is
presented in figure 6. In the standard PIFA the surface current has a
maximum distribution near the DC-shorting strip. In the proposed antenna
similar behavior can be observed. The high concentrated surface current
path lengths represent the resonating strip lengths responsible for resonance
at various frequencies. Table 1 shows the gain value of the proposed
antenna at various resonating frequencies.

IJCSBI.ORG
Table 1. Gain of proposed antenna at resonant frequencies
Frequency (GHz)
parameter
0.8 1.5 2.7
Gain (dBi) 1.469 2.872 5.428
4. CONCLUSION
The presented compact PIFA design has enough frequency bandwidth to
support GSM, LTE, and DCS bands having directivity ranging from 1.4dBi
to 5.4dBi. The proposed antenna has an appreciable return loss at different
resonant frequencies 0.8GHz, 1.5GHz and 2.7GHz. The low profile flat
design of presented multiband PIFA for cellular user equipment is simple
and easy to fabricate. Matching circuit used in the modal has given fine
tuning possibility making it to resonate at different frequencies. The
proposed antenna is recommended for smart phones with multiband
functionality.
5. ACKNOWLEDGMENTS
The authors are pleased to acknowledge the support of management of the
organizations for providing all required facilities to carry out the work, and
sincere thanks to principal for suggestions.
6. FUTURE SCOPE
This modal can be further improved to have more frequency bands like
LTE700, LTE2300 etc., and matching circuit can be optimized for better
profile advantage.
REFERENCES
[1] F-H Chu and Kin-Lu W, Planar printed strip monopole with a closely-coupled
parasitic shorted strip for eight-band LTE/GSM/UMTS mobile phone, IEEE
Transactions. on Antennas and Prop., V. 58, No. 10, Oct. 2010, pp. 34263431.
[2] Keng-Chih Lin, Chih-Hao Lin and Yi-Cheng Lin, Simple printed multiband antenna
with novel parasitic-element design for multi standard mobile phone applications,
IEEE Trans. on Antennas and Propag., Vol. 61, No. 1, Jan. 2013, pp. 488491.
[3] Yong-Ling Ban, Jin-Hua Chen, Joshua Le-Wei Li and Yujiang Wu, Small-size
printed coupled-fed antenna for eight-band LTE/GSM/UMTS wireless wide area
network operation in an internal mobile handset, IET Microw. Antennas Propag., Vol.
7, Iss. 6, Jun. 2013, pp. 399407.
[4] Y. K. Park and Y. Sung, A reconfigurable antenna for quad-band mobile handset
applications, IEEE Trans. on Antennas and Propag., Vol. 60, No. 6, Jun. 2012, pp.
30033006.
[5] Yun-Wen Chi, Kin-Lu Wong, Quarter-wavelength printed loop antenna with an
internal printed matching circuit for GSM/DCS/PCS/UMTS operation in the mobile
phone, IEEE Trans. on Antennas and Propag., Vol. 57, No. 9, Sept. 2009, pp. 2541
2547.

IJCSBI.ORG
[6] Jae Hee Kim, Won Woo Cho and Wee Sang Park, A small dual-band inverted-F
antenna with a twisted line, IEEE Antennas and Wirel. Propag. Lett., Vol. 8, 2009,
pp. 307310.
[7] Chien-Wen Chiu and Yu-Jen Chi, Planar hexa-band inverted-F antenna for portable
device applications, IEEE Antennas Wirel. Propag. Lett., Vol. 8, 2009, pp. 1099
1102.
[8] Sung. Y, Compact quad-band reconfigurable antenna for mobile phone applications,
Electron. Lett., Vol. 48, No. 16, 2012, pp. 977979.
[9] N. Misran, M. M. Yunus and M. T. Islam, Small Dual-Band Planar Antenna with
Folded Patch Feed, Journal of Applied Sciences Research, Vol. 6, No. 12, pp. 1975-
1980, 2010.
[10] Mikko Komulainen, Markus Berg, Heli Jantunen, Eekki T. Salonen and Charles Free,
A Frequency Tuning Method for a Planar Inverted-F Antenna, IEEE Trans. on
Antennas and Propagat., Vol. 56, No. 4, Apr. 2008, pp. 944-950.
[11] Z. D. Liu, P. S. Hall and D. Wake, Dual-Frequency Planar Inverted-F Antenna, IEEE
Trans. Antennas Propagat., Vol. 45, No. 10, Oct. 1997, pp. 1451-1458.
[12] F. R. Hsiao, H. T. Chen, T. W. Chiou, G. Y. Lee and K. L. Wong, A Dual-Band
Planar Inverted-F Patch Antenna with a Branch-Slit, Microwave and Optical
Technology Letters, Vol. 32, No. 4, Feb. 2002, pp. 310-312.
[13] Sandeep Kumar Veeravalli, K Shambavi, Zachariah C Alex, Design of tri band
antenna for mobile applications International conference on Communication and
Signal Processing, April 3-5, 2013, pp. 947-950.

Jackson, S. J. and Thomas, T., 2016. Enhanced-bandwidth Compact Printed
Inverted F Antenna Suitable for LTE/GSM Mobile Phone. International
Journal of Computer Science and Business Informatics, Vol. 16, No. 1, pp.
24-31.

IJCSBI.ORG
Implementation of e-Government
Services using Java Technologies:
A Methodological Guide in
Improving an e-Police Management
System in Zimbabwe
Mahlangu Gilbert, Chikonye Musafare, Furusa Samuel Simbarashe
and Mugoniwa Beauty
Computer Science and Information Systems

Midlands State University
Gweru, Zimbabwe
ABSTRACT
In recent years, there has been a paradigm shift in service delivery in government
ministries, departments and agencies. Many governments have redefined their service
delivery systems through the implementation of e-government using various technologies.
This has seen the transfer of major government services to online platforms. The purpose of
this study was to appraise the implementation of e-government services using Java
technologies by focusing on an e-police management system in Zimbabwe so that
Zimbabwe Republic Police (ZRP) can align its systems with other National Police Services
world-over. Using a qualitative research method, the researchers collected data from the
participants through qualitative questionnaires, semi-structured interviews and focus group
discussions. The findings indicated that the majority of services that are offered by the ZRP
are not electronically enabled. The study proposes that ZRP can improve its service
delivery system by utilising Java technologies presented in this study. This will enable the
organisation to solve the problems that are inherent in the current system.
Keywords
E-government services, Java technologies, e-police management system
1. INTRODUCTION
In recent years, there has been a paradigm shift in service delivery in
government ministries, departments and agencies. This has been fostered by
the introduction of e-government systems around the world. Governments
have entered into an intense competition of occupying a leading position in

IJCSBI.ORG
the e-government arena (Bilbao et al., 2013). Inspired by Denmark and

Estonia who have played a leading role in implementing e-government
(OECD, 2011), many governments have redefined their service delivery
systems by using various technologies to expand their services online (Al-
khouri, 2013). Technology plays an important role in transforming
government services. It has become a key component in managing and
administering public affairs. The use of Information Communication
Technology (ICT) have been seen as a catalyst in improving the
performance and delivery system of government ministries, departments
and agencies (OECD, 2009). ICT has the prospective of supporting
governments to bond people, government departments, suppliers and
service partners to achieve greater efficiencies and public value. However,
OECD (2009) argued that the adoption and use of e-government services
remains low and unsatisfactory in other countries, despite the fact that
various ICTs that exist in the world of technology have the potential to drive
the implementation of e-government.
Using various policies and legislatures, governments world-over have
invested heavily in programmes that enable the transfer of major
government services to online platforms. Similarly, in Zimbabwe the
government directed its ministries, departments and agencies to implement
e-government systems in 2013 as enshrined in the ZIMASSET policy
document (Government of the Republic of Zimbabwe, 2013). This was
aimed at improving public service delivery and sharing of critical
information that can enhance efficiency among the arms of the government
and its citizens through the use appropriate technologies. The main purpose
of this study was to appraise the implementation of e-government services
using Java technologies by focusing on an e-police management system in
Zimbabwe so that Zimbabwe Republic Police (ZRP) can align its systems
with other National Police Services world-over.
While technology has become a key driver for the implementation of e-
government globally, there is a non-appearance of explicitly synchronised
efforts at government ministries and departments in Zimbabwe to transform
government services to e-government (Cunningham and Cunningham,
2008). E-government services in Zimbabwe are characterised by
disintegration and discrepancies which has caused inconsistencies in service
delivery. Correspondingly, the current ZRPs service delivery system is yet
to provide its major services online. In spite of the fact that some of the
internal systems are computerised, electronic delivery of service is yet to be
implemented. Furthermore, there are no integrations between ZRP systems
with other government departments and agencies whom they are supposed
to share with critical data. For instance, the verification of fingerprints

IJCSBI.ORG
which has to be done with the registrar generals national database. In

addition, vehicle clearance for the purpose of vehicle registration requires
verification with ZRPs database of stolen vehicles. Moreover, applications
that could allow the general public to make police reports that do not require
the citizen to visit a police station are not implemented. Complainants have
to visit police stations to report a crime. The use of electronic forms and
Short Message Services (SMS) would be very convenient for the public to
conduct such reports in the comfort of their homes. With the scenarios
presented above, it can be noted that there is a very huge communication
barrier between the police department, citizens, other government
departments and agencies in Zimbabwe. Hence, it is important that suitable
technologies are used to solve these problems.
The major objective of this study was to provide a methodological guide on

improving an e-police management system in Zimbabwe through the
application of Java technologies. To achieve this objective, the following
sub objectives have been formulated:
a) To analyse the state of e-government services at ZRP
b) To establish services that are electronically enabled at ZRP
c) To identify areas in ZRPs systems where e-government should be
implemented using Java technologies to allow the general public to
access police services online.
2. CONCEPT OF ELECTRONIC GOVERNMENT (E-

GOVERNMENT)
E-government is characterized by the presence of a two way communication
between the government agencies and the citizens, and a total integration of
the entire government services for online accessibility (Singh, 2015). It is a
system that is implemented using digital means in order to provide a single
gateway to various governmental services. Therefore, the aspect of e-
government denotes the provision of various governmental services using
ICTs, mainly the internet.
Various organisations and authors have provided different definitions on e-
government. The World Bank (2012) defines e-government as the use of
ICTs such as wide area network, the internet and mobile computing in
improving service delivery among citizens, business organisations and other
government departments and agencies. As defined by Devasena & Balraj
(2014), it is the implementation and delivery of government services
through ICTs to achieve efficiency, effectiveness, transparency,
accountability and information sharing. Guided by the purpose of the
research, this study will define e-government as the transfer of government

IJCSBI.ORG
services to online platforms using web and mobile technologies to facilitate

online and remote access.
2.1. E- government services

Various services exist within the domain of the e-government. Some of the
services that signifies the existence of e-government include: e-elections, e-
police, and e-healthcare, e- banking, e-tax declarations, e-school, citizen
self-service portals, government portals and e-land (Sharma, 2010).
However, the success of these services depends on the level of adoption of
ICTs among the government ministries, departments and agencies.
Therefore, for e-government to be an effective proxy in the provision of
government services, it must be implemented using appropriate
technologies that support mobility and remote access.
2.2. E-police management system as an e-government service

E-police management system is an integrated platform that uses web and
mobile applications to address the operational needs of all units of a
National Police Force (NPF). These include crime management,
investigations, automated biometric identification, forensic analysis & data
exchange within the police force and other government departments and
agencies. According to Chavan et al. (2014), an e-police is a government
service that uses ICTs brokering systems in increasing the professional
efficiency for the government police administration. It is an e-government
service that uses ICTs to facilitate communication between the police
department and the citizens of a country in order to improve administrative,
operational and professional efficiency. Using an e-police management
system, police officers can find crime related information from citizens,
other police stations and national databases for further action. However, the
success of this system will depend upon the development of an interactive
website and an electronic database.
3. JAVA TECHNOLOGY
Java consist of various technologies which are subsidiary elements of the
three main suites; Java Enterprise Edition (J2EE), Java Standard Edition
(J2SE) and Java Mobile Edition (JME). These technologies can support
embedded, web, stand alone, distributed and mobile application
development. The utilisation of these technologies can allow the
development of different types of Java application that can be used in the
implementation of e-government services. In this section, the subsidiary
elements of Java technologies that are used for web and mobile application
development are presented.

IJCSBI.ORG
3.1. Java Servlets

Java Servlets is a Java based web technology used by web developers to
create efficient web solutions that are accessed through the request-response
programming model. It is the most used Java technology for developing
dynamic web pages. Servlets are also known as server side programs
because they services the HTTP requests from the client-side, process them
and returns back the HTTP responses. They extend the functionality of the
server by enabling a connection between the client and the server
(Saraevi, 2011). They act as a middle layer between requests from web
browsers and databases on HTTP servers. This technology enables the
server to process multiple requests at once using the threading model;
thereby improving performance. According to Qian (2007), the Servlet
technology has seen Java expanding greatly in most enterprise web
application development. Most enterprise information systems that support
web services are developed using the Servlet technology.
3.2. Java Server Pages (JSP)

JSP technology function with the combination of HTML and Java in order
to provide the dynamic component of the Web. It enables programmers to
mix regular and static pages with dynamically generated content using
special tags (Bergsten, 2002). The elements of the JSP technology
determines how the page builds dynamic content. Like the servlets, JSP
enables users to interact with the databases and web services, and also
process responses according to client request. The technology is also
platform and server independent because it can run on different operating
systems or web servers. While there are various technologies that are used
to build web applications that serve dynamic content, JSP has really caught
the attention of the web development community (Koletzke, Dorsey &
Faderman, 2003). The process of creating JSP accessible on the web is
much simpler than other web development technologies because it does not
require packaging of the program.
3.3. Enterprise Java Beans (EJB)

EJB technology addresses the need to support distributed, transactional,
secure and portable applications based on Java technology (Kumaran,
2002). It is a framework for the server-side of enterprise Java applications
that support database driven websites. The use of EJB in developing web
applications makes Java a distributed technology which can be accessed
from different servers and platforms. This technology support database
programming by providing services like transaction management and
persistence management (Stevens, 2010). The technology was developed
for the purpose of reducing repetitive work involved in persistence,

IJCSBI.ORG
transaction management and database security. As opposed to servlets, the

EJB is a pure business component that support load balancing, clustering,
resource pooling and caching (Dolgicer et al., 2003).
3.4. Java Data Base Connectivity (JDBC) API

The JDBC API defines interfaces and classes for writing database
applications in Java. JDBC provides a standard library for connecting to
relational databases using Structured Query Language (SQL) with exactly
the same Java syntax. However, for the connection to be established the
java application must call the JDBC library so that it loads the driver that
provides a link to the database (Manonaniam, 2000). It can be used in
servlet, JSP or EJB technologies to access the database when implementing
database driven websites. Therefore, the main function of JDBC is to
establish connections to databases, provide access to a given database,
provides mechanisms for reading, inserting, updating and deleting entries of
data in a database and takes care of transactions composed of different SQL
statements (Stevens, 2010).
3.5. Mobility Pack

Mobile applications are created for mobile use through Java technology.
The applications can run on devices like cellular phones, tablets, PDAs and
palmtops to mention but a few. Java has a dual role in the development of
mobile applications. First, the application should support the sending and
receiving of SMS, whether single or bulk between two ends or devices.
Secondly, it should support the development of applications that are
compatible with mobile devices that uses different operating systems. The
two roles of Java in mobile application development can be implemented
using mobility pack. This is a Java API that enables software developers to
integrate SMS technology into their solutions and helps them develop
platform independent applications suitable for mobile devices (El-Kassas et
al., 2015).
4. METHODOLOGY
A qualitative research method based on qualitative questionnaires and semi-
structured interviews were conducted to ascertain whether ZRP is providing
some of its services online to the general public. These research instruments
also helped the researchers to identify areas of application for Java
technology in implementing an e-police management system. Focus group
discussion was used to critical evaluate the methodological guide for the use
of Java technologies in improving an e-police management system in
Zimbabwe. The participants for completing the qualitative questionnaire
were randomly selected from the general public. Although the selection for

IJCSBI.ORG
questionnaire was random, the participants were supposed to have reported

a case before. This helped the researchers to establish the state of the e-
government services at ZRP. As for the semi-structured interviews,
participants were selected using a stratified sampling technique from all the
operational units of ZRP at the 2015 Provincial Agricultural show.
According Hair et al. (2015), this sampling technique enables the researcher
to achieve a proportional representation from the study population. The
sample population for focus group discussions was composed of 12 certified
Java programmers from Zimbabwe.
5. FINDINGS
This section presents the findings of the study based on the information
gathered from the police officers and the citizens of Zimbabwe.
5.1. The state of e-government at ZRP

In this subsection, the state of e-government at ZRP is presented in view of
the services offered through web or mobile technologies.
5.1.1. Citizens using web or mobile technology to report crime

In this category, all the respondents indicated that there is no website at
ZRP that allows crimes to be reported using this facility. Crimes are
reported physical at the police station or by calling the police station. The
mobile technology is limited to phone calls only. Although phone calls are
part of the facility to report a crime, it was noted that in some instances the
line(s) may be continuously engaged or some police officers are not willing
to record crime reports from phone calls.
5.1.2. Citizens using web or mobile technology to send suggestions

There are no electronic suggestion boxes at ZRP stations in Zimbabwe. All
the respondents indicated that they are only able to send suggestions to the
police using WhatsApp platforms. Each police station has a WhatsApp
number that can be used by the citizens to send suggestions and tip offs.
5.1.3. Citizens using web or mobile technology to track progress on their

reports
The citizens are not able to use web or mobile technologies to track the
progress of their reports. They are supposed to visit the police station where
the case was reported.
5.1.4. Citizens accessing notices via the web or mobile technology

IJCSBI.ORG
Notices that include policies and procedures that support law enforcement
are not accessible by either web or mobile technology. In most cases,
citizens are caught by surprise by the law enforcement agents.
5.1.5. Police officers using web or mobile technology to clear vehicles and
finger prints
Clearances for vehicle and finger prints are done manually. The details of
the vehicle or the person to be cleared are captured at the police station and
sent through manual dispatches to the Head Office which has the central
database. The officers at the Head Office verify the records as per clearance
required. In turn, the clearing officer action the clearance documents
accordingly and send them back to the dispatching office using the same
delivery channel.
5.1.6. Police officers using web or mobile technology to verify drivers

licences
Verification of drivers licences for possession and genuine has been a major
challenge since it is enforced at roadblocks which is some distance away
from the police station. In most cases, motorist are delayed at the
roadblocks for failure to produce a drivers licence or for being suspected to
be holding a fake drivers licence. This is because police officers at the road
block do not have the facility to conduct instant verification. The motorist
has to be taken to the police station or have the vehicle impounded until the
drivers licence has been produced or verified with Central Vehicle Registry
to be genuine through a phone call.
5.2. Services that are electronically enabled at ZRP

The success of an e-government service depends on the number of services
that are electronically enabled. The findings of this study noted that the
majority of services that are offered by the ZRP are not electronically
enabled. Citizens need to visit or call the department in order to access a
service. It was found that the only service that is electronically enabled is
crime reporting through WhatsApp platform. Each police station has a
WhatsApp number that can be used by the citizens to report crime or give
suggestions.
6. APPLICATION AREAS OF JAVA TECHNOLOGY FOR E-

POLICE MANAGEMENT SYSTEM
Java technologies can be utilised in different areas of e-government and for
various e-government applications that can be implemented by ZRP. To
ensure that an e-police management system is implemented in Zimbabwe, in

IJCSBI.ORG
this section areas of application for e-police management system where java
technologies can be used are identified and explained.
6.1. E-crime reporting

This service is expected to provide online crime reporting to citizens using
web or mobile technologies. Any citizen who is a victim of crime can send a
report to the nearest police station by completing a web form or using a
predefined format for SMS. The web forms may be developed using JSP or
servlet while an SMS gateway may be developed using Java eclipse with
auto-response functionality for sending delivery reports to the crime
reporter. For the crimes to be routed to specific tables, JDBC API can be
utilised to handle the INSERT query to the crime management database.
6.2. E-suggestion
This service will allow citizens to send their suggestions that can assist the
police officers in tracking criminals and detecting crime. An electronic
suggestion box need to be implemented using JSP and JDBC API. JSP is
necessary for developing web forms for inputting suggestions while JDBC
is needed for allowing the insertion of suggestions to the electronic
suggestion box and retrieving them for further analysis. The major SQL
statements for this service are the INSERT and SELECT.
6.3. E-clearance
This service will provide a linkage between the police department and
government departments that has the central databases for vehicle and
fingerprints. There is need for system integration especially between ZRP's
Stolen Motor Vehicle Database and CVR's vehicle registration system to
allow verification of vehicle before registration by CVR. In addition,
integration is needed between ZRP's Fingerprint Identification System (FIS)
and National Registration System to allow verification of accused identity
before creation a criminal record of same by ZRP. The JDBC and the EJB
API can be used in this service since the two APIs are used to provide
connection among relational databases and handling database transactions
respectively. This linkage will unlock government-to-government e-
government.
6.4. E-crime progress tracking

This service will allow complainants and the police officers In-Charge of
crime investigations at various stations to check the productivity of the
investigating officers using web technologies and the intranet respectively.
To implement this service, JSP will be required for creating the login
function while in turn JDBC will handle database queries.

IJCSBI.ORG
6.5. E-notices
This service will enable the police to transfer policies and procedures to the
online platform, giving easy access of information to citizens instead of
placing policies and procedures on notice boards at the charge office. Other
critical notices may include the names of criminals on the wanted list,
warrant of arrests, missing persons, people detained at police stations, lost
and found property and victims of road accidents. This service can be
implemented using JSP or Java Servlet to provide dynamic web content.
Apart from putting notices on the website, crime alerts can also be executed
as SMS using Ozeki Message Server and JDBC, whereby the SELECT
statement written in Java will be used to query the database table for
outgoing messages to the citizens.
6.6. E-drivers license verification

This service will enable the police officers at the roadblock to verify the
existence of a drivers license from CVR using mobile technology while
those at the police station may use web technology. To implement the
mobile functionality, a powerful SMS solution can be developed with an
auto reply facility. This can be done using an SMS gateway created in Java
eclipse, database triggers or stored procedures enforced using JDBC while
supported by the INSERT and the SELECT statements.
6.7. E-payment
The presence of this service to the e-police management system is expected
to allow citizens to pay for services and fines that are administered by ZRP
using available electronic or mobile payment systems in Zimbabwe. For
example, fingerprint vetting and road traffic offenses. In this regard, the JSP
is needed for developing form based and API based gateways to provide
authentication, authorisation and integration. The other technologies for this
service include the mobility pack which enables the development of
applications for mobile devices and the JDBC for handling payment records
of services and fines.
7. FURTHER RESEARCH
The study looked at the basic and core business of the police force in
Zimbabwe. In order to fully utilise an e-police management system, there is
need to incorporate the aspect of intelligent system in order to increase
crime control. Furthermore, future research may also look at how the
technology may be used to handle the administrative needs of the police
force. For example, e-procurement and e-transfers of police officers.

IJCSBI.ORG
8. CONCLUSION
Java technologies have come a long way in developing mobile and web
technologies. Most applications that are needed to allow online service
delivery can be developed using Java. The use of Java technologies for the
implementation of e-police management system by ZRP is expected to
improve the quality of services, reduce the costs of delivering services,
improve the utilisation of scarce resources, enhance accountability and
transparency and restore citizen confidence in government services. This
study noted that there are a number of services ZRP can provide online to
Zimbabwe citizens and its officers. For the aforesaid reason, ZRP need to
implement e-government system to conform to other National Police
Services world-over. Hence, this study suggested a methodological guide of
providing an e-police management system anchored in a web portal
developed using Java technologies that are capable of supporting the core
business of the police force in Zimbabwe.
9. ACKNOWLEDGEMENTS
The authors are grateful to ZRP officers, the citizens of Zimbabwe and Java
programmers who participated in this research. The information they
provided was vital for the success of this research.
10. REFERENCES
Al-khouri, A.M., 2013. e-Government in Arab Countries: A 6-Staged Roadmap to Develop
the Public Sector. , 4(1), pp.80107.
Bergsten, H., 2002. JavaServer Pages,
Bilbao, B., Dutta, S. & Lanvin, B., 2013. The Global Information Technology Report 2013:
Growth and Jobs in a Hyperconnected World, Available at:
http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:The+Global+Infor
mation+Technology+Report+2013+Growth+and+Jobs+in+a+Hyperconnected+Worl
d#1.
Chavan, R. et al., Securing Information Brokering In Distributed Information Sharing
Using RC6 and SPEKE. , 2(2).
Devasena. C.L. & Balraj, P.L., 2014. E-Governance in Southern States of India - Towards
Whole-of-Government ( WoG ) Abstract: , 3(2).
Dolgicer, B.M., Bayer, G. & Bardash, M., 2003. Java Servlets and Enterprise Java Beans In
Enterprise Architectures: Friends or Foes Part I.
El-Kassas, W.S. et al., 2015. Taxonomy of Cross-Platform Mobile Applications
Development Approaches. Ain Shams Engineering Journal. Available at:
http://linkinghub.elsevier.com/retrieve/pii/S2090447915001276.
Hair, J.F. et al., 2015. Essentials of Business Research Methods, Routledge. Available at:
https://books.google.com/books?id=GfnqBgAAQBAJ&pgis=1 [Accessed December
21, 2015].

IJCSBI.ORG
Kumaran, S.I., 2002. Jini Technology: An Overview, Prentice Hall PTR. Available at:
https://books.google.com/books?id=WtxQAAAAMAAJ&pgis=1 [Accessed
December 21, 2015].
Manonaniam, 2000. Advanced Java Programming with Database Application.
OECD, 2009. OECD e-Government Studies Rethinking e-Government Services User-
Centred Approaches: User-Centred Approaches, OECD Publishing. Available at:
https://books.google.com/books?id=Og7WAgAAQBAJ&pgis=1 [Accessed
November 24, 2015].
OECD, 2011. OECD Public Governance Reviews Estonia: Towards a Single Government
Approach, OECD Publishing. Available at:
https://books.google.com/books?id=n2HWAgAAQBAJ&pgis=1 [Accessed
December 14, 2015].
Qian, K., 2007. Java Web Development Illuminated, Jones & Bartlett Learning. Available
at: https://books.google.com/books?id=oY9fShrQyUgC&pgis=1 [Accessed
December 14, 2015].
Saraevi, M., 2011. Integrated Management Web- Based Applications on Java Platform.
Quality, pp.573580. Available at: http://muzafers.uninp.edu.rs/refer.html.
Singh, S., 2015. E-Governance State-of-the-Art Survey: Stuttgart , Germany. , pp.4765.
Stevens, P., 2010. Web Development in Java Warning More information HTTP basic
Layers Terminology that s not specific to Java. Agenda.

Gilbert, M., Musafare, C., Simbarashe, F. S., Beauty, M. 2016.
Implementation of e-Government Services using Java Technologies: A
Methodological Guide in Improving an e-Police Management System in
Zimbabwe. International Journal of Computer Science and Business
Informatics, Vol. 16, No. 1, pp. 32-43.

IJCSBI.ORG
Onto Tagger: Ontology Focused

Image Tagging System
Incorporating Semantic Deviation
Computing and Strategic Set
Expansion
Gerard Deepak
Department of Computer Science and Engineering
University Visvesvaraya College of Engineering
Bangalore University, Bangalore, India.
Sheeba Priyadarshini J
Department of Computer Science
St. Josephs College (Autonomous)
Langford Road, Bangalore, India
ABSTRACT
Social Tagging of images uploaded to the Web is highly mandatory as tags serve as the
entities for image retrieval. Manual Tagging of images makes the overall process tedious
and moreover the tags when manually assigned become noisy. Several automatic tag
recommendation systems are available but the background study proves that the tag
relevance is not very high. In the era of Semantic Web, there is a need for a semantic driven
tagger which would perform efficiently. Also, a system which bridges the gap between
manual and automatic tag recommendation is required. An ontology driven semantic tagger
for tagging images with social importance which tags the images based on limited reference
tags is proposed. The proposed methodology combines ontology crawling using K-Means
Clustering and Semantic Deviation Computation using Modified Normalized Google
Distance Measured. The tag space is enhanced using Strategic Set Expansion incorporating
a dynamic semantic deviation computation. An average precision percentage of 84.4 and an
F-Measure percentage of 86.67 are achieved.
Keywords
Ontology Tagger, Semantic Deviation, Social Tagging, Strategic Set Expansion, Tag
Recommendation.
1. INTRODUCTION
The World Wide Web is a repository of billions of images, video, text and
other data. Owing to the increasing number of users of the World Wide
Web, even the data on the Web is increasing exponentially. With the
popularity of trends in Social Networking, the amount of social data is also

IJCSBI.ORG
increasing almost every minute. It has been estimated that Social data
mainly includes images and videos that are shared among the social actors
(users and their friends) when compared to text and other kinds of data.
Images are among the topmost in the list of social data that are the most
shared.
The problem targeted is definitely not the increasing amount of social or
web data or their means to handle it but retrieving the exact or highly
similar items like images or video when searched is the main issue. Tags or
Annotations become a very important means to retrieve web multimedia
information. To facilitate easy searching or retrieval of exact images or
items from the World Wide Web, Tagging of such social data items plays an
important strategy as images without tags over the Web are
ubiquitous[1].There can be two approaches which can be followed. The first
approach is the traditional approach where the Search Engine or The Search
Algorithm must be designed in a more efficient or intelligent manner to
extract the required items. The second of the two approaches is by correctly
tagging the social items like images and reducing the noise in such tags such
that there is a quicker and faster convergence to retrieve the exact images or
videos from the World Wide Web. However, there exists diversity in social
tagging behaviors [2].
Tagging an image or a video or any such social items like blogs is like
specifying the right address for the search engine to retrieve the tagged item.
Tags bridge the gap between users cognition and objective [3].Tagging that
exists in the present day scenario varies from manual to automatic tag
recommendation. Tagging an image on the social websites is a methodology
of organizing the images and reliable tags must be specified for social
photos [4]. Several automatic taggers are available but the best tag
recommendation system must be semantically driven and must recommend
them by re-organizing the tags such that the users may be convinced by the
tags that are recommended by the search engine. Also the semantics of the
tags assigned must correlate with that of the Web Search Engine such that
the items that are browsed must have high relevance and must be
significant.
Although, with the availability of several approaches for automatic tag
recommendation, a semantic web approach which is actually budding for
tagging will make it much efficient to comply with the web search engines.
An ontological approach is proposed for tag construction and
recommendation in order to minimize and remove noisy tags for images of
community importance. Ontologies constitute important intrinsic structures
of the Semantic Web. Due to a paradigm shift towards improved semantic
web technologies, ontology focused approach for image tagging is a highly

IJCSBI.ORG
commendable and feasible approach. The usage of ontologies towards

tagging images not just enhances the relevance of the results but also
increases the correctness and the quality of tags. The tags when included
based on the ontologies, the probability of deviations and noise is
minimized to a very large extent. Images with community importance are
mainly concentrated because such images are searched more frequently by a
large number of people and most of the times simultaneously. There are
several social networking sites specifically for uploading of images like
Flickr, Instagram, etc. Henceforth, a tagger that computes similarity of tags
and automatically recommends by lowering tag noises is mandatory.
Motivation: The absence of a tagger which semantically tags images of
community importance to without redundancy is the intrinsic motivation of
developing such a system. A tagger which tags images by automatically
recommending tags which similar images are also holding in order to reduce
the tag redundancy. A tagging system which is semantically driven is in a
high demand in the era of intelligent and semantic Web. The main motive of
the proposed system is higher the quality of the tags better is the relevance
of the images. Thus, there is a need for a system that automatically
recommends tags which are just not straight forward but are of a very high
quality by aggregating tags the tag space with relevant images. Most of the
users who upload images feel that manual tagging is much better than
automatic tagging. Thus, a gap is formed between the approach of manual
tagging and automatic tagging that needs to be bridged. The Proposed
approach fills this gap and is highly proficient when tagging images.
Contribution: A system that is highly efficient in tag recommendation
which semantically computes for similar tags based on user driven reference
tag is proposed. An ontological approach which uses tag level and image
level ontologies for tagging is proposed. An innovative approach which
incorporates K-Means Clustering for extracting similar images and
eventually their tags is proposed for constructing the tag space.
Furthermore, the ontologies in the tag space are re-ranked based on semantic
similarity deviation computed using Modified Normalized Google Distance
[5] [6] is implemented. Dynamic computation of semantic deviation is used
for checking the tag level semantic similarity. Several Permutations of the
tags from this tree are obtained and are recommended to the users for
selection of tags.
Organization: The paper organization is as follows. The Section 2 provides
a brief overview of related research work. Section 3 describes the Problem
Definition. Section 4 presents the Proposed System Architecture. Section 5
discusses the implementation in brief. Results are presented in Section 6.

IJCSBI.ORG
Performance is evaluated in Section 7.Finally, Section 8 concludes the

paper.
2. RELATED WORKS
Zhou et al., [8] have proposed a hybrid collaborative model for image
tagging which incorporates probabilistic techniques with a content based
strategy for image tagging. The strategy uses an image tag associative
matrix for and non-negative factorization of matrix for collaborative
filtering. Correlations are estimated using a norm method. The approach is
tested using three large databases with a large number of images and tags.
Pantraki et al., [9] have proposed an image tagging and recommendation
system which uses the strategy of Parallel Strategy Analysis 2 in which
three matrices are used with visual features, user information and the tag
information. This strategy was tested with ample data sets. The major
drawback of this technique is that it uses three large data structures which
may tend to increase the overall complexity of the solution.
Zhu et al., [10] have proposed an approach where the social information of
the image is used as social clues for tagging of images. Along with social
clues, the social group information and the tagging preferences of the image
owner is used to predict and annotate tags. A strategy called as
neighborhood voting on the Online Social Networks is used. This approach
is quite good but the tag space doesnt predict numerous tags which
becomes a disadvantage in this system. Qian et al., [11] have proposed a
novel strategy of incorporating diversified semantics for image tagging and
annotation over online social networks. The strategy uses considering the
factor of relevance of improving the tag quality. The usage of semantic
compensation onto the already determined top ranked tags is one of the
major concepts on which this strategy is based.
Liu et al., [12] have proposed an innovative methodology of using more
than a single for tagging of an image. This multi feature tagging approach
learns several scores of features which are further converted to weights. A
tag pair matrix of a predormintarily low rank is formulated such that several
features are a significant in tagging of an image. Lin et al., [13] have
proposed a strategy of image tagging on social websites by extracting sparse
patterns of tags from tags that are already available in the web data. The tags
which the users have contributed already are being used to predict and re-
recommend tags in this approach. The problem of minimizing the quadratic
loss is very well treated in this strategy by incorporating a bi-layer norm.
Though there would be a good level of tag relevance, there will be a
dependency on the already available tags rather than the image level
compatibility.

IJCSBI.ORG
Im et al., [14] have proposed a novel strategy where semantics is

incorporated. The semantic relationship between a pair of tags is inferred in
this approach through a semantic paradigm. The context of the word that is
used to tag an image is strategically found out in this approach. The
disadvantage is that the contexts might add a little amount of noise into the
final predicted tags. Pliakos et al., [15] have proposed a unique
methodology of image tagging as well as geological tag inference
simultaneously. The underlying strategy is the usage of hypergraph incident
matrix for geographical location prediction. A concept of group sparsity
constraints enforcement is incorporated to this methodology to get better
results.
Barmpoutis et al., [16] have proposed an approach in which the strategy of
tensor structure decomposition of tags. This is purely a mathematical
approach which aims at breaking the links between the multilink relations of
the tag elements. This method involves the usage of four different unique
matrices with four innovative mathematical strategies for tag
decomposition. Pavlidis [17] has studied the interdependence of image
tagging with that of the actual content of the image. This paper clearly
points out that the complexity involved in the analysis of the contents of the
image makes tagging and annotation of images also complicated.
Furthermore, the paper proposed the non-pixel dependent strategies for
image tagging rather than the conventional methodologies.
Huges et al., [18] have proposed an approach which incorporates machine
learning techniques for computing and estimating the tag relevance of
images that are tagged based on Geo-Spatial Information. The tags are
classified using an SVM classifier and a semantic approach is used for
computing the tag heterogeneity. This approach reduces the irrelevant tags
and improves the overall tag quality and is applicable for geographically
tagged images of community importance. The approach combines a
statistical geographical distribution for spatial information estimation and
Machine Learning Technique like SVM for tag classification. Wang et al.,
[19] has proposed a strategy which bridges the gap between the features
extracted and the overall context. An approach of context regularization is
implemented for the visual features which makes this strategy an effective
one for image tagging over social networks.
Gong et al., [20] have proposed a robust model of image annotation using
the strategic computation of semantic similarity between words. A language
specific model is designed for estimating the semantic similarity in this
approach. A cross media relevance model and a translation model is
integrated with this strategy in order to enhance the relevance of the results
in this approach. Sawant et al., [21] have proposed an innovative approach

IJCSBI.ORG
of social image tagging based on social inputs. Online Social Websites are a
repository for very large volumes of social data like the features and patterns
of users, images and social tags. This approach mainly concentrates on the
semantic analysis of these social inputs to formulate and predict tags for
annotation.
3. PROBLEM DEFINITION
Tagging of images is of most importance as the tag names play a vital role
when the images are retrieved in the search engine. The major objectives of
our work which were tackled as individual problems are:
To automatically recommend tags for an image uploaded driven by a
limited input tag of users reference.
To implement an ontological approach for tag recommendation.
To overcome and avoid noisy and redundant tags.
To improve the overall quality of the tags for community important
images.
To improve the recall, precision, f-measure and accuracy of the
proposed system.
4. PROPOSED SYSTEM ARCHITECTURE
Figure 1. Proposed System Architecture

IJCSBI.ORG
The overall architecture of the proposed system is depicted in Figure 1

where in a user initially uploads an image on the web through a social
networking site or directly through the Application Programming Interface
(API) of the system created. The API of the proposed system has its own
Graphical User Interface (GUI) and the user can directly interact through the
systems API with the World Wide Web. The system developed also has a
facility to integrate its GUI with that of social networking sites like
Instagram or Flicker where images that are of community interest are
uploaded. The only restriction for integrating the systems API with that of
social networks is that the user must have an account in the social
networking site to which the user wants to post his image through the
proposed system.
As the user is uploading the image, it is a mandate that the user also
specifies ten reference tag names which the user wishes to tag. The
reference image tags are parsed and tokenized to remove redundancy and
extract individual tag elements. The individual tag elements are further used
as input query to cluster similar images. Furthermore, the tags of the
resultant images are extracted and semantic similarity is computed to that of
the reference tags. If semantic similarity is high, the tags are included in the
tag space. The tag space is generally a vector which is dynamically
constructed where in the tags which are highly similar are organized and can
be used in several instances.
To enhance the relevance of the tags and to give many options, the
ontologies similar to the reference tags are crawled using a focused crawler
and once again the semantic similarity is computed with that of the
reference tags and is included into the tag space. The focused crawler is
integrated into the system which automatically crawls the ontologies of
similar tags based on the input reference tags. Based on the weights of
semantic similarity index, re-ranking of tags is done. Furthermore, using a
combinatorial function, the tags are predicted to the user. The user can
finalize the tags that can be used. The proposed system enhances the tag
relevance and minimizes redundant and noisy tags. This ensures a very high
tag quality. Moreover, since ontology is also considered as a parameter for
tag formulation, the relevance of the tags is very high.
5. IMPLEMENTATION
The implementation of the proposed system is done using JAVA as a
programming language and Netbeans as the IDE. The reference tag names
are initially parsed and tokenized. Tokenization is performed by
incorporating a JAVA based NLTK (Natural Language Tool Kit) Tokenizer

IJCSBI.ORG
to split the individual tag elements. Also a redundancy checker which is

custom written in JAVA is used for removing redundant tag items.
Redundancy of tag elements is checked to minimize the complexity of the
system and avoid unnecessary checks for duplicate items. This definitely
enhances the performance of the system implemented.
The reference tag elements are used to cluster similar images using a
standard nearest neighbor approach incorporating a basic K-Means
Clustering. Furthermore, the semantic similarity computation using
Modified Google Distance Measure is carried out. A Focused Ontology
Crawler is integrated into the system which facilitates the phenomenon of
extracting ontologies nearest to the reference tag items. As the ontologies
are crawled, a fitness function incorporating Modified Normalized Google
Distance which dynamically computes semantic similarity of the ontologies
and checks with that of the reference tag items is also imbibed into the
system. The motive behind this fitness function is to retain the ontologies
which are correlative to the reference tags and enhance the overall relevance
of the tags and accelerate the quality of tags. Also, the ontology information
of the image uploaded is fed into the tag space for tag recommendation.
A tag space is formulated which is a vector containing all the semantically
similar tags and as well as the best fit ontologies. The tag space is further
enhanced by applying the strategic set expansion technique. Further, the re-
ranking of the tags is done based on semantic weights of the tags. The
combinatorial function is subject to the tags in order to predict the tags for
finalization. The combinatorial function implemented is based on a neural
network model with initial training such that the tags are predicted as per the
users choice. Parameters like the size of the tags, number of words needed
to the tag, etc., serve as dynamic inputs to the adaptive artificial neural
network for prediction.
Strategic Set Expansion refers to a set expansion paradigm where deviations
are computed among the existing sets in the Tag Space. The standard
threshold for the deviation is considered as average of the sum of initial
deviations in the tag space before the set expansion is applied. During set
expansion, the key procedure which takes place is when a new term is
included; the semantic weight of the term is computed and is checked with
the threshold deviation. If its lesser than the threshold, then it is included in
the set. Otherwise it is discarded if it exceeds the threshold. The immediate
neighborhood with an order of maximum two hierarchies of the tags in the
tag space is alone considered for set expansion. If TS be the current Tag
Space, then the elements of the tag space are listed in (1). Deviation
between a pair of tags in the tag space is computed and the Deviation set is
depicted by (2).

IJCSBI.ORG
TS= {ts1, ts2, ts3, ts4..} (1)

D= {d1, d2, d3, d4.} (2)
Where di=Semantic Difference (di, di+1), such that i= {1, 2, 3.}.
If Td is the threshold of the deviation, then Td = D/n where n is the number
of elements in the tag space. In order to compute the Semantic Similarity
Deviation or the Semantic Heterogeneity, a Modified Normalized Google
Distance is considered which is given by the (3).x and y in (3) signifies the
terms whose semantic heterogeneity has to be determined. The Google
Distance Measure is modified owing to a reason that it yields better and
more precise results increasing the overall confidence of the proposed
approach.
max{log f ( x), log f ( y)} log f ( x, y) log( x y) log( x y) (3)

ModfNGD( x, y)
logN log(x y) - logx - logy - min{log f(x), logf(y)}
5.1 EXPERIMENTATION
The experiment was carried out by using 1124 images which were of
community importance. 350 images were collected from personal
photographers while the rest were crawled from Bing and Google Image
search engines. Several monuments which are popular and the ones which
are not very popular are also included for experimentation. The
amalgamation of less popular monuments with that of the popular ones
actually makes the system ready for every kind photographs of community
importance.
5.2 ALGORITHM
The Proposed Algorithm which inputs a reference tag pair and predicts tags
is discussed in Table 1.
Table 1: The Proposed Tag Recommendation Algorithm

Begin
Step 1: Initialize the reference tag pair (rs1, rs2) for a specific image that is
to be uploaded. The reference tags are incorporated as a discrete tag pair set
RS. Ensure at least 1 tag pair is input for an image.
Step 2: The reference tag set RS is sent into a parser and then a Tokenizer
to extract elements, remove redundancy and remove stem words and further
a new set RSfin is formulated.

IJCSBI.ORG
Step 3: If RSfin is non-empty, For each element do

a) Cluster Semantically Similar Images
b) Extract the individual tags from these Images and dynamically compute
the semantic similarity of the tags using Modified Normalized Google
Distance Measure. The tags are tokenized and stemmed to remove
redundancies.
c) Construct a Tag Space vector with the new tags TS and apply strategic
set expansion technique
Step 4: Crawl the essential ontologies O1, O2. On corresponding to the
reference tags and Check for assertions in the ontologies.
Step 5: While the ontologies are not assertions, compute the semantic
similarity of the ontologies and check with that of the reference tags. Add
them into the tag space TS.
Step 6: Using strategic set expansion, expand the Tag Space by
Dynamically Computing the Semantic Heterogeneity.
Step 7: Re-rank the tags based on the Semantic Measure.
Step 8: Using Combinatorial Function, Predict and Recommend the tags to
the user.
End
6. RESULTS
The tagging for several images that are uploaded are governed by the system
implemented. Several images that are of community importance and are of
interest to others are incorporated into the experiments. Various images that
were crawled from the Google and Bing image search engines and were
further uploaded through the proposed system for tagging. Table 2 depicts
the images and their corresponding tags which were recommended by the
system. The tags recommended were checked for correctness and relevance.
The tags recommended by the system must be acceptable by the user and
the user must finalize the tags based on the systems recommendation. The
proposed system predicts and recommends tags of a high quality which will
be evaluated in the performance evaluation

IJCSBI.ORG
Table 2: Relevant Tags Recommended for images of social relevance

Images Uploaded Reference Tags Recommended
Tags
Amber Fort Amber Fort, Jaipur
Amber Fort, Jaipuri
Amber Fort, Amer Fort,
Amber Amer Fort, Amer
Palace, Amer Fort Palace,
Amber Palace Fort,
Amber Royal Fort, Royal
Amer Palace, Rajasthani
Amber Fort, Rajasthan
Fort Amber, Amber Lake
Fort.
Bom Jesus Basilica of Bom Jesus,
Church, Goa Parish Church of Bom
Jesus, Borea Jezuchi
Bajilika, Baslica do Bom
Jesus,Francis Xavier
Church, Goa Xavier
Church, Goa Bom Jesus
Church, Relic of St.
Francis Xavier Church,
Basilica of the Relic of
Francis Xavier, Old Goa
Church, Holy Jesus
Church, Good Jesus
Basilica.
Philomena Saint Philomena Church,
Church, Philomena,Church
Mysore. Mysore, Mysore Church,
St. Josephs Cathedral
Mysore, Mysore
Philomena Josephs
Church, Mysore Ashoka
Road Parish Church.

IJCSBI.ORG
Velankanni Our Lady of Good Health

Church Basilica, Mother Marys
Shrine Velankanni,
Velankanni Basilica,
Annai Velankanni
Aalayam, Arokiamarie
Basilica Velankanni,
Arokiya Matha Church.
Fateh Prakash Fateh Palace, Prakash
Palace Palace, Royal Lake
Palace, Lake Palace
Uadipur, Prakash Lake
Palace, Udaipur City
Palace Complex, Lake
Palace Complex, Pichola
Lake Palace, Pichola
Prakash Palace, Pichola
Fateh Palace.
Brihadeshwara Thanjai Periya Koyil,
Temple Thanjavur Periya Kovil,
Tanjavur Brihadeshwara Temple
Tanjore, Thanjavur,
Brihadeeswarar Temple
Thanjavoor, Thanjai
Kalvettu Koil,
Peruvudaiyar Kovil,
RajaRajeswara Temple,
Rajarajeswaram.
7. PERFORMANCE EVALUATION
The performance analysis of Onto Tagger is carried out by using four
evaluation metrics namely the Recall, Precision, F-measure and Accuracy
which is depicted in Equations (4), (5), (6) and (7) respectively. The
percentage of the metrics used is considered in the approach proposed. The
Precision or Positive Predictive Value referred to as a Fraction of Retrieved
Instances [22] and is incorporated as the ratio of the number of relevant
recommended tags to that of the tags formulated in the Tag Space. Recall
also called as sensitivity is known as the fraction of relevant instances [22]

IJCSBI.ORG
and is incorporated as the ratio of relevant recommended tags to that of the

total number of relevant tags. F-Measure or F-Score relates Recall and
Precision and is known as the harmonic mean [22] of Recall and Precision.
Accuracy is formulated as the average of recall and precision.
Number of Relevant Tags Recommended

Precision (4)
Total number of Tags Formulated in Tag Space
Number of relevant Tags Recommended (5)

Recall
Total number of Relevant Tags
2 * Precision * Recall
F - Measure (6)
Precision Recall
Precision Recall
Accuracy (7)
2
The images which were considered for performance evaluation are Amber
Fort, Bom Jesus Church Goa, Philomenas Church Mysore, Velankanni
Church and Qutab Minar. The images chosen range from the most popular to
the less famous images. All images chosen have an importance to a specific
community or a society in general. The percentage of Recall, Precision, and
Accuracy for the images is individually depicted in Table 2. Also, the
average measures of the depicted evaluation metrics is computed which is
also lists in Table 2.
Table 2 : Performance Measurement of Onto Tagger

Image Recall Precision Accuracy
% % %
Amber Fort 88 83 85.5
Bom Jesus 93 88 90.5
Chrurch,Goa
Philomenas Church, 86 75 80.5
Mysore
Velankanni Church 92 86 89
Qutab Minar 87 90 88.5
Average 89.2 84.4 86.8

IJCSBI.ORG
Onto Tagger is compared with Tag Relevance [7] which is used as a bench
mark for performance comparison. The Precision Percentage of Onto Tagger
was compared with that of Tag Relevance. Five arbitrary Precision values of
Tag Relevance were considered which was then converted into a percentage
of precision for comparison with the Precision Percentage of Onto Tagger.
Figure 2 gives the graphical depiction of the Precision Percentage of Onto
Tagger and Tag relevance. It is clearly evident from the line graph that the
Performance of Onto Tagger is much better than that of Tag Relevance.
Figure 2: Comparison of the Precision Percentage of Onto Tagger with Tag Relevance
The graphical depiction of the F-Measure of the individual image elements

considered for performance evaluation is depicted in Figure 3. The F-
Measure of Amber Fort and Bom Jesus Church Goa is 85.43 % and 90.43 %
respectively.Philomenas Church, Velankanni Church and Qutab Minar have
F-Measure Value of 80.12 %, 88.89 % and 88.47 % respectively. The
Comparison of the Average Precision Percentage of Onto Tagger and Tag
Relevance is depicted in Figure 4 as a bar chart. It is clearly evident that
Onto Tagger outperforms Tag Relevance by 1.9 %. From this we can infer
that Onto Tagger is 1.9 percent more precise than Tag Relevance.

IJCSBI.ORG
Figure 3: F-Measure in Percentage for various images
Figure 4 : Comparison of Average Percentage of Precision of Onto Tagger with that of

Tag Relevance

IJCSBI.ORG
8. CONCLUSIONS
An innovative approach for automatic tag recommendation which is based
on limited reference tags is implemented successfully. The paradigm is
driven by ontology extraction and incorporates a Modified Normalized
Google Distance measure to compute semantic deviations between ontology
entities and tags. A strategic set expansion strategy is incorporated to
enhance the population of the tag space. The proposed methodology clusters
similar ontologies using a basic K-Means Clustering Algorithm. Several
community contributed images that are crawled from Flicker and Google
Image Search are used for validating the approach. Experimental evidences
show that the Proposed onto Tagger approach is the best-in-class method for
tagging social images on the Web. Onto Tagger yields an average precision
of 84.4 %, an average recall of 89.2 %, accuracy of 86.8 %. The average
proposed F-measure of Onto Tagger is 86.67 %. Onto Tagger incorporates a
strategic set expansion technique which increases the overall number of tags
recommended. Onto Tagger extracts the ontology information of the similar
tags. Owing to the reason of incorporating ontological information, the tags
are of high relevance. Dynamic Semantic Deviation Computation that has
been used in Onto Tagger eliminates the noisy and irrelevant tags. Onto
Tagger is one of the image tag recommendation systems which is driven by
the reference input tags which makes it a cognitive bridge between manual
tagging and automating tagging systems.
ACKNOWLEDGMENTS
I whole heartedly thank my parents who stood by me and supported me
during the implementation of this work. Above all, I thank God the
Almighty and Everlasting Father and my Lord Jesus Christ who gave me
strength, wisdom and health for the completion of this work.
REFERENCES
[1] Deng, Z.H., Yu, H. and Yang, Y., 2015, October. Image Tagging via Cross-Modal
Semantic Mapping. In Proceedings of the 23rd Annual ACM Conference on
Multimedia Conference (pp. 1143-1146). ACM.
[2] Liu, Q. and Li, Z., 2016. Projective nonnegative matrix factorization for social image
retrieval. Neurocomputing, 172, pp.19-26.
[3] Jabeen, F., Khusro, S., Majid, A. and Rauf, A., 2016. Semantics discovery in social
tagging systems: A review. Multimedia Tools and Applications,75(1), pp.573-605.

IJCSBI.ORG
[4] Hong, Y., Chen, T., Zhang, K. and Sun, L., 2016, January. Personalized Annotation for
Mobile Photos Based on Users Social Circle. In MultiMedia Modeling (pp. 76-87).
Springer International Publishing.
[5] Wikipedia article https://en.wikipedia.org/wiki/Normalized_Google_distance last
accessed on Feb 29th, 2016.
[6] R.L. Cilibrasi and P.M.B. Vitanyi, The Google similarity distance , IEEE Trans.
Knowledge and Data Engineering, 19:3(2007), 370383 or
http://arxiv.org/abs/cs.CL/0412098
[7] Mishra, D., Singh, U.P. and Richhariya, V., 2014. Tag Relevance for Social Image
Retrieval in Accordance with Neighbor Voting Algorithm. IJCSNS, 14(7), p.50.
[8] Zhou, N., Cheung, W.K., Qiu, G. and Xue, X., 2011. A hybrid probabilistic model for
unified collaborative and content-based image tagging. Pattern Analysis and Machine
Intelligence, IEEE Transactions on, 33(7), pp.1281-1294.
[9] Pantraki, E. and Kotropoulos, C., 2015, September. Automatic image tagging and
recommendation via PARAFAC2. In Machine Learning for Signal Processing
(MLSP), 2015 IEEE 25th International Workshop on (pp. 1-6). IEEE
[10] Zhu, S., Aloufi, S. and El Saddik, A., 2015, June. Utilizing image social clues for
automated image tagging. In Multimedia and Expo (ICME), 2015 IEEE International
Conference on (pp. 1-6). IEEE.
[11] Qian, X., Hua, X.S., Tang, Y.Y. and Mei, T., 2014. Social image tagging with diverse
semantics. Cybernetics, IEEE Transactions on, 44(12), pp.2493-2508.
[12] Liu, X., Liu, R., Cao, Q. and Li, F., 2013, November. Multi-feature Late Fusion for
Image Tagging. In Pattern Recognition (ACPR), 2013 2nd IAPR Asian Conference
on (pp. 34-37). IEEE.
[13] Lin, J., Yuan, J., Duan, L.Y., Luo, S. and Gao, W., 2012, July. Social Image Tagging
by Mining Sparse Tag Patterns from Auxiliary Data. In Multimedia and Expo (ICME),
2012 IEEE International Conference on (pp. 7-12). IEEE.
[14] Im, D.H. and Park, G.D., 2013, June. STAG: semantic image annotation using
relationships between tags. In Information Science and Applications (ICISA), 2013
International Conference on (pp. 1-2). IEEE.
[15] Pliakos, K. and Kotropoulos, C., 2014, May. Simultaneous image tagging and geo-
location prediction within hypergraph ranking framework. InAcoustics, Speech and
Signal Processing (ICASSP), 2014 IEEE International Conference on (pp. 6894-6898).
IEEE.
[16] Barmpoutis, P., Kotropoulos, C. and Pliakos, K., 2015, September. Image tag
recommendation based on novel tensor structures and their decompositions. In Image
and Signal Processing and Analysis (ISPA), 2015 9th International Symposium on (pp.
7-12). IEEE.
[17] Pavlidis, T., 2009, June. Why meaningful automatic tagging of images is very hard.
In Multimedia and Expo, 2009. ICME 2009. IEEE International Conference on (pp.
1432-1435). IEEE.
[18] Hughes, M., Connor, N.E. and Jones, G.J., 2012, May. A machine learning approach to
determining tag relevance in geotagged Flickr imagery. In Image Analysis for
Multimedia Interactive Services (WIAMIS), 2012 13th International Workshop on (pp.
1-4). IEEE.

IJCSBI.ORG
[19] Wang, L., Zhao, Z., Su, F. and Sun, W., 2014, July. Content-based social image
retrieval with context regularization. In Multimedia and Expo Workshops (ICMEW),
2014 IEEE International Conference on (pp. 1-6). IEEE.
[20] Gong, T., Li, S. and Tan, C.L., 2010, October. A semantic similarity language model to
improve automatic image annotation. In Tools with Artificial Intelligence (ICTAI),
2010 22nd IEEE International Conference on (Vol. 1, pp. 197-203). IEEE.
[21] Sawant, N., Li, J. and Wang, J.Z., 2011. Automatic image semantic interpretation using
social action and tagging data. Multimedia Tools and Applications, 51(1), pp.213-246.
[22] Wikipedia article https://en.wikipedia.org/wiki/Precision_and_recall as accessed on
Feb 29th, 2016.

Gerard, D. and Sheeba, P., J., 2016. Onto Tagger: Ontology Focused Image
Tagging System Incorporating Semantic Deviation Computing and Strategic
Set Expansion. International Journal of Computer Science and Business
Informatics, Vol. 16, No. 1, pp. 44-61.

Vol 16 No 1 - January-June 2016

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Vol 16 No 1 - January-June 2016

Caricato da

Copyright:

Formati disponibili

International Journal of Computer Science

and Business Informatics

ISSN: 1694-2507 (Print)

Various Common Control Channel Assignment Schemes in CRNs: A Survey ......................................... 1

Implementation of e-Government Services using Java Technologies: A Methodological Guide in

Various Common Control Channel

Inderdeep Kaur Aulakh

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 1

2. VARIOUS COMMON CONTROL CHANNEL ASSIGNMENT

Fig. 1 Categorization of CCC design [7].

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 2

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 3

3.1 Dynamic CCC assignment using Swarm Intelligence

3.2 Cluster based CCC assignment

3.3 Hybrid CCC based Media Access Control Protocol

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 4

3.4 Synchronized MAC Protocol

3.5 Sequence based Rendezvous Common Control Channel Design

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 5

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 6

This paper may be cited as:

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 7

Framework for Threat Modelling

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 8

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 9

Therefore power utilities need to model threat to the prepaid electricity

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 10

4. OVERVIEW OF THREAT MODELLING

5. PRESENT APPROACHES TO THREAT MODELLING

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 11

5.1 Attack Tree-Model

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 12

Figure 1: The OCTAVE framework

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 13

5.3 Microsoft Security Development Lifecycle (SDL) model

5.3.1 Threat overview

5.3.2 Attack strategies in AMI

6. PROCESSES/STEPS IN THREAT MODELING

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 14

Figure 2: Process in threat modelling

6.1 Identify Assets

6.2 Create an architecture overview

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 15

6.2.2 Identify the threats.

6.2.3 Document the threats.

6.2.4 Rate the threats.

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 16

8. FINDINGS AND DISCUSSION

8.2 Vulnerabilities, threats and attack strategies

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 17

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 18

Defeat electricity billing/enjoy free electricity consumption

Physical bypass Cyber-attack Mechanical induced

Partial bypass Wholly bypass Microcontroller prog. Infra-red prog. Mechanically

Figure 3: Outline of intrusion scenarios for prepaid electricity metering

10. PROPOSED FRAMEWORK

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 19

Phase 1 Check for lead information: Brainstorm (internal/external):

Outline intrusion/attack scenarios: Describe attack goal (s):

Define & profile potential attackers: Rate threats: (PIVE)

Mitigation: Define protection strategy:

Figure 4: Proposed framework

ISSN: 1694-2108 | Vol. 16, No. 1. JANUARY-MARCH 2016 20

13. FUTURE RESEARCH