Please use this identifier to cite or link to this item:
Title: Constructing a web application for identification, detection and analysis of pattern based fraud
Authors: Mavrova, Aneliya Dicheva
Keywords: Electronic commerce
Data mining
Machine learning
Bayesian statistical decision theory
Web applications
Issue Date: 2017
Abstract: The rapid development and popularization of the e-commerce business in the past few decades have made online purchasing of products preferable and more comfortable shopping method for buyers. As credit / debit cards became most common payment method for both online and regular purchases, the number of registered fraudulent transactions also increased significantly. Fraudsters seem to find back-doors in the security systems, leading to accumulation of losses for banks, merchants and customers. Therefore implementation of efficient fraud detection system has been made imperative for bank issuing systems in order to be able to predict suspicious behavior and stop fraud before it occurs. Modern techniques for fraud prevention include Machine Learning, Data Mining, use of Artificial Immune Systems etc. Based on the dataset and the present information various techniques could be used to develop security systems. This project has a goal to examine the transaction process for both genuine and fraudulent purchases of online products and using a Naive Bayes algorithm to be able to detect an flag the suspicious such. The algorithms selection is based on the dataset size (small dataset, since credit card details include sensitive data) and due to the high probability performance of Naive Bayes when data is not clustered or is partially missing.
Description: B.SC.IT(HONS)
Appears in Collections:Dissertations - FacICT - 2017

Files in This Item:
File Description SizeFormat 
  Restricted Access
3.64 MBAdobe PDFView/Open Request a copy

Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.