Sei sulla pagina 1di 6

Data Mining Journal Article Review J Layman

MG520.XG1 Unit 4 Assignment A- Journal Article Review Jacquelyn Layman Globe University Minneapolis

Data Mining Journal Article Review J Layman

Abstract In this paper I will summarize and review the Journal of Information Systems article entitled Data Mining of Electronic Mail and Auditing: A Research Agenda (Debreceny, Gray, 2011, pg 195-226). This article discusses the technique of mining data from corporate emails known as Data Mining of Electronic Mail, or DME, as a source of audit evidence as was the case in the Enron Corporation scandal.

Overview Of The Data Mining Of Electronic Mail (DME)

Data Mining Journal Article Review J Layman

Electronic mail is the preferred method of communication in corporate America today, both internally and externally with customers, suppliers, business partners and the like. Debrency and Gray refer to email as tangible evidence of the enterprises communication DNA and a vital form of audit evidence (2011, pg. 196). The date and time stamp on emails, along with detailed header information including sending server ip address and receiving server ip address make emails an indisputable paper trail of communication, minus the paper. Corporations automatically copy to back-up all in and outbound email data in the event that there is a power or server failure and simply deleting these emails does not remove them from the backed up files server side, it merely deletes them from the local system. This information in cases of audit has been mined to date using either the directed or ex post approach which mines only information meeting the criteria of being between certain dates, or certain individuals and more recently an undirected or ex ante approach has been introduced using less stringent and more flexible searching methods such as topic or unusual content similar to the algorithms used to sort spam from wanted emails. An increasing use of the mined data is in the realm of fraud protection. Although mining of data from electronic communications is relatively new to the mainstream, the legal profession has been using this method of discovery for some time. With the increased reliance on electronic communication in litigation, the Federal Rules of Civil Procedure have been updated to extend the requirements of defendants in litigation to identify and produce this electronic evidence and as such enterprises are creating policies and processes for the long-term retention of these server back-ups for potential use in the audit and legal processes.

Implementations

Data Mining Journal Article Review J Layman

The auditing of electronic communications has been implemented in the arena of financial statement fraud detection and prevention, as a forensic investigation tool into motive, or intent in litigation, and as an important tool in reducing regulatory and compliance risks. Using these methods, potentially unwanted behavior can be predicted and put to an end before it becomes a security risk. This type of analysis seeks to find potentially incriminating words or phrases, predicting deception by uncovering changes in writing style or tone, an increase in the number of emails to or from key accounts, introduction of new themes or subjects into emails in key departments or groups or changes in the communications between participants in key groups or departments with parties from outside the organization. Email Data Mining Techniques Email data mining combines traditional data mining with structured analysis of the dissected sections of the email including the header which consists of the sender, the recipients the route information, time and date stamp and the subject, and the body which contains unstructured data in a text field. Text mining the email body The email body and email attachments are unstructured text and as such standard text mining techniques can be implemented. Text mining is an undirected form of data mining that is designed to seek out and identify unknown patterns in unstructured textual data. This data mining takes on two separate forms; categorization/classification identifies themes either with the use of a thesaurus or by recognizing differences and similarities. Concept linkage identifies blocks of text and linking them to like concepts. Text mining is currently being used in other domains including marketing, accounting, spam identification, and public policy. Email body

Data Mining Journal Article Review J Layman

text searches can include keyword searches, or deception searches in either content or identity. Analysis of the email header The email header contains structured information that can be easily mined and analyzed. The analysis of the email header can include volume and velocity, social network analysis or internal and external recipient analysis. Summary Data mining of electronic mail can be a useful source of information in the auditing and litigation processes. The data collected through keyword searches, pattern searches or searches of volume and velocity can be used as evidence in the audit or in the courtroom. This evidence is undeniable, and is signed with a date and time stamp that shows the originating and receiving server addresses. These emails are stored on a large server, where they are backed up in case of system loss. These backups provide insight into the content of material being shared, as well as into the character of the person sending and receiving the data. This type of insight becomes highly beneficial in managing corporate security, maintaining healthy relationships with partners and customers, and early detection of fraudulent or illegal behaviors. Maintaining a reputation among peers and customers is crucial to the success of a company, as is ensuring the safety of secret corporate information. The internal audit process can detect the early warning signs that this trust will be breached. DME can be an effective part of this audit process.

Data Mining Journal Article Review J Layman

References Debreceny, Gray, (2011). Data Mining of Electronic Mail and Auditing:A Research Agenda. Journal of Information Systems. 25 (2), pp.195-226

Potrebbero piacerti anche