Sei sulla pagina 1di 5

Application of Data Mining in Predicting Employee’s

Performance Analysis using Decision Tree Algorithm


Blances M. Sanchez
Polytechnic University of the Philippines Graduate School
0648-A Purok 7 PNR Site FTI Compound, Western Bicutan
Taguig City
+639771005895
blancessanchez30@gmail.com

redefining the industry, be it in terms of technology or research.


ABSTRACT Data mining techniques is used to extract the information and
The ability to predict the success of an employee in the could be used to improve the existing scenarios. We can also
industrial sector is very significant. Almost every organization find existing relationships and patterns. Data Mining combines
has a performance evaluation of its employees. The performance machine learning, statistics and visualization techniques to
of the employee is based on various factors such as personal, discover and extract knowledge. Employee’s retention has
social, abilities, punctuality, efficiency, interpersonal become an indication of industrial performance and enrollment
relationship, and task/responsibility, quality of duty, actual management. Here, potential problem will be identified as
results, psychological and other environmental variables. In this earlier.
paper we aim to develop a system that would allow the
management to conclude the employee’s performance based
from gathered data. For now, the scope of this work limits only There are many classification techniques in Data Mining such as
to data collection from surveys. A similar procedure can be Decision Tree, Naïve Bayes Classifier, Nearest Neighbor,
performed to extract information from employees in the future. Support Vector Machines, etc. Decision Tree is among the
Data mining techniques provide useful information from the popular classification techniques, which can produce the logic
historical projects depending on which the management can statement. The generated information can use for future
make decisions for high-quality workforce. Our proposed references. Data Mining is being used by many because of its
system would also take employee data such as the attendance, tools with potential to forecast and can be a contribution to
pay rate, revenue, expense, gender and remark to analyze the knowledge that could help better in decision making.
performance of employees of the organization. Our proposed
system would be using Decision Tree Algorithm for evaluating
the performance of the employees. The main objective of this paper is to propose a system that to
monitor the employee’s performance in a fair, objective and
Keywords comprehensive manner. The propose system can be used for
Employee Performance, Employee Performance Prediction, extraction of knowledge for predicting employees’ performance.
Employee Performance Analysis, Decision Tree Algorithm, Based on the employee’s performance results possible to take
Data Mining Classification decision whether advanced training, talent enrichment or further
qualification required or not. Though there have been many
previous studies in this domain, there have been issues that still
1. INTRODUCTION need to be addressed. We are trying to build and use this
research to build a better more reliable model that cannot be
Most of the organization or companies have a formal generalized to different scenarios but also function well on
performance evaluation system in which employee job various data sets with varying properties with minimal or no
performance is graded on a regular basis, usually once or twice a adjustments.
year. A good performance evaluation system can prominently
benefit an organization. It helps employee behavior toward
organizational aims by permitting employees know what is
expected for them, and it yields information for making 2. METHODS
employment decisions, such as those regarding pay raises, 2.1 Research Questions
promotion or releases. An employee can improve their
performance by way of monitoring the progression of their Listed below is the set of research questions addressed in the study
performance [2]. 1. The meaning of decision tree algorithm and how will
it process data in order to get employee’s analytic
results.
Data Mining is the process of analyzing data from different 2. The variables that the organization will be used as basis
perspectives and summarizing it into useful information [1]. It is for filtering needed data.
also sometimes called knowledge discovery in databases (KDD)
[2]. Data mining is the next big revolutionary field that is
Variables Possible Values
1. Quality Good, Bad.
2. Productivity Meets expectations,
Improvement Needed.
3. Independence Meets expectations,
Improvement Needed.
4. Reliability Meets expectations,
Figure 1. Framework for employee’s performance Improvement Needed.
analyzation process 5. Job Skills Serious, Common.
6. Interpersonal Meets expectations,
relationships Improvement Needed.
2.2 Research Design 7. Cooperation Meets expectations,
The employee performance data will be collected by the Improvement Needed.
proposed system. There will be two modules in the propose 8. Commitment Meets expectations,
system that will be used in order to get interpretative employee Improvement Needed.
performance result. 9. Attendance Meets expectations,
Improvement Needed.
a. Module that will be used by employee’s personal
10. Initiative Perfect, Ordinary.
computers to get data while working
11. Creativity Meets expectations,
b. Module that will process that collected data in
Improvement Needed.
order to come up with analysis of employee’s
performance 12. Adherence to Meets expectations,
the Policy Improvement Needed.
13. Overall Meets expectations,
2.3 Data Understanding Performance Improvement Needed.
The data that will be used in the proposal will be collected by
the sub-module of the propose system. This sub-module will be Table 1. Predictive Variables
created and installed in every employee’s devices to monitor
their activity during work hours. This will be applicable to A. Quality – The extent to which an employee’s work is
industry where using computer is the main equipment in the completed thoroughly and correctly following established
organization. process & procedures. Required paperwork is thorough and neat.

2.4 Data Collection B. Productivity / Independence / Reliability – The extent to


which an employee produces a significant volume of work
Data will be collected by using survey obtaining performance efficiently in a specified period of time. Ability to work
information about employees from the Project-leads and independently with little or no direction/ follow-up to complete
Managers. tasks / job assignment.
a. Generation of data source of predictive variable by
means of survey from the system that will be C. Job Knowledge - The extent to which an employee possesses
developed and demonstrates an understating of the work instructions,
b. Identification of different factors, which affects an processes, equipment and materials required to perform the job.
employee’s performance during industrial career Employee possesses the practical and technical knowledge
c. Construction of a decision tree model using required of the job.
classification data mining techniques on the basis of
identified predictive variables and their values
D. Interpersonal Relationships / Cooperation / Commitment –
The extent to which employee is willing and demonstrates the
2.5 Data Mining Process ability to cooperate, work and communicate with coworkers,
Data are analyzed using classification method to predict the supervisors, subordinates and/or outside contacts. Employee
employee performance. accepts and responds to change in a positive manner. Accepts
job assignments and additional duties willingly, takes
a. Data Preparation responsibility for own performance and job assignments.
The data set will be used, which will be obtained from
chosen organization’s various departments. Data will
be stored in different tables. E. Attendance – The extent, to which an employee is punctual,
observes prescribed work break/meal periods and has an
b. Data Selection and transformation acceptable overall attendance record. Employee’s willingness to
Fields will be selected, that is required for data work overtime as required.
mining. All predictive values will be selected. All the
predictor and response variables which were derived
from the database are given in Table 1. F. Initiative/ Creativity – The extent to which an employee seeks
out new assignments, proposes improved work methods,
suggests ideas to eliminate waste, finds new and better ways of
doing things.
G. Adherence to Policy – The extent to which the employees First phase: Creation of forms that has definite number of
follows organization policies, procedures and work conduct questions and all are 1 mark each. Data is collected from the
rules. Complies with and follows all safety rules and regulations, responses from the forms given by the employee of the
wears required safety equipment. organization.

H. Lead (if applicable) – The extent to which the employee Second phase: The analysis is to be made on the collected
demonstrates proper judgment and decision-making skills when responses of the employees. The responses are collected by
directing others. Directs work flow in assigned areas effectively means of the proposed system. The response for every test is
to meet production / area goals. maintained in the database. Features are extracted from the
collected data i.e. the names and the scores are collected.
I. Overall Performance – Rate employee’s overall performance
in comparison to position duties and responsibilities.
Third phase: The employees will be rated on their performance
in the survey conducted. Applying the decision tree would make
help in making decision making.
2.6 Mining Process
To Construct the Decision Tree, we use following method
1. Select a variable of training samples as nodes; create a
branch to every possible value of the variables.
Accordingly, the training sample set is divided into
several sub-sets.
2. Do the same method to each branch, Training sample
is the subset corresponding to the branches and one of
the subsets which its parent node is divided into.
When the node of all the training samples belongs to
the same classification, or no remaining attributes can Figure 2. System Architecture of getting employee
be used to further divide, Or the branch does not have information
samples, stop splitting the node branching and make it
a leaf node.

Decision tree is a classifier in the form of a tree structure where


each node is either:
 A leaf node - indicates the value of the target attribute
 A decision node - specifies some test to be carried out
on a single attribute- value, with one branch and sub-
tree for each possible outcome of the test.

2.7. Proposed System


Traditional approaches were implemented for such systems
with the data collection through offline tests, marks at the time
of interview process and their domain knowledge [3].
Limitations must overcome if will implement an online system
for assessing the employees. The proposed approach is to assess Figure 3. System Architecture of Proposed System
the performance of the employees by driving knowledge of them
from all the domains/perspectives.
3. RESULTS
The employee will be judged on the basis of various diverse Data mining consists a set of techniques that can be used to
factors such as personal, social, abilities, punctuality, efficiency, extract relevant and interesting knowledge from the data. It has
interpersonal relationship, task/responsibility, quality of duty, several tasks such as association rule mining, classification and
actual results, psychological and other environmental variables. prediction and clustering. Decision tree generates a decision tree
Managers, subordinates, peers and employee itself answer from the given training data.
questions about an individual’s skill and attributes. To avoid the
difficulty in precisely assessing the performance of each
employee, it is necessary to precisely quantifying the score of
each candidate. 4. DISCUSSION & CONCLUSION
This paper proposes a system that uses data mining and decision
tree as a tool to gather employee’s performance data and makes
There are three phases: decision. We use decision tree algorithm to classify employee
Initial phase: Selection and preparation of questionnaire. easily and take appropriate decision quickly. Further actions can
be taken in this circumstance to avoid any danger related to Productivity for an Organization.
hiring poorly performed employee. Future work involves more
proper data from several companies. When the model is [14] Lamarca, B., & Ambat, S. (2017) The Development of a
generated, these algorithms could be developed for predicting Performance Appraisal System Using Decision Tree Analysis
performance of employees in any kind of organization. and Fuzzy Logic.

[15] Muralidhar, A., Kumar, S., Modi, A. (2017) Employee


ACKNOWLEDGEMENT Performance Appraisal System based on Ranking and Reviews.
I would like to express my gratitude to Dr. Rosicar Escober and [16] Zadgoankar, A. S., Singh, J. & Agrawal, R. (2016)
to Dr. Benilda Eleonor Commendador for their insightful Summative Assessment for Performance Evaluation of Faculty
comments and encouragement. Their guidance helped me in all using Data Mining Techniques.
time of research and writing. I could not have imagined having a
better advisors for my Masters study. [17] Vaddepalli, M., & Rajesh, A. (2017) Data Mining:
Evaluating Employee's attendance attribute of an educational
institute using classification algorithm based on decision tree.
REFERENCES
[1] Kirimi, J., & Moturi C. (2016) Application of Data Mining [18] Baykara, B. (2015) Impact of Evaluation Methods on
Classification in Employee Performance Prediction. Decision Tree Accuracy.

[2] Sarker, A., Shamsim S.M., Zama S., & Rahman M. (2018) [19] Chikte, P. & Nikose, A. (2016) Employee Performance
Employee's Performance Analysis and Prediction using K Evaluation using Genetic Algorithm.
Means Clustering & Decision Tree Algorithm.
[20] Quan, P., Liu, Y., Zhang, T., & et al. (2018) A Novel Data
[3] Chungade, T., & Kharat, S. (2017) Employee Performance Mining Approach Towards Human Resource Performance
Assessment in Virtual Organization using Domain-Driven Data Appraisal.
Mining and Sentiment Analysis

[4] Gupta, S., Gupta, A., & Thakur, G. (2015) Data Mining for
Prediction of Human Performance Capability in the Software
Industry.

[5] Srinavasan, R., Vishnavat, K., & Singh, H. (2018) Employee


Performance and Leave Management using Data Mining
Technique.

[6] Zhao, Y., Hryniewicki, M., & Cheng F. (2018) Employee


Turnover Prediction with Machine Learning: A Reliable
Approach.

[7] Goyal, A., & Sharma, M. (2015) An Application of Data


Mining to Improve Personnel Performance Evaluation in Higher
Education Sector in India.

[8] Sikaroudi, A., Ghousi, R., & Sikaroudi A. (2015) A data


mining approach to employee turnover prediction (case study:
Arak automative parts manufacturing).

[9] Kanwal, Z., Noor, W., & Jaffar, Z. (2019) Predictive Human
Resource Analytics Using Data Mining Classification
Techniques.

[10] Deo, P., Uplenchwar, D., Mahajan, C., & Tupe, S. (2017)
Employee Performance Evaluation System using ID3
Algorithm.

[11] Gao, Y. (2017) Using Decision Tree to Analyze the


Turnover of Employees.

[12] Ameen, Z., & Mahmood A. (2018) Employee Performance


Assessment using Modified Decision Tree.

[13] Yousuf, M. & Karthick, J. (2016) Predicting Human

Potrebbero piacerti anche