ABSTRACT
Fraud, waste, and abuse in many nancial systems are estimated to result in sig-ni cant losses annually. Predictive analytics o er government and private nancial institutions the opportunity to identify, prevent or recover such losses. This work proposed a novel Big Data driven approach for fraud detection based on Deep Learn-ing methods. A supervised Deep Learning solution leveraging Big Data was shown to be an e ective Fraud predictor. Additionally, an unsupervised method based on anomaly detection using deep autoencoders was proposed for when there is few or no labelled data. The two methods presented o ered adaptive and predictive Fraud de-tection through improved Analytics. Future work will look into how the two methods can be integrated into an e ective tool for enhanced Fraud detection.
TABLE OF CONTENTS
Abstract
List of Tables
List of Figures
CHAPTER ONE
1 Introduction
1.1 Background to the Study
1.2 Problem Statement
1.3 Aim of the Study
1.4 Significance of the Study
1.5 Scope of the Work
1.6 Structure of the Report
CHAPTER TWO
2 Literature Review
2.1 Frauds and Fraud detection
2.2 Rule-based fraud detection
2.3 Statistical fraud detection
2.4 Machine Learning Fraud detection
2.5 Deep Learning
CHAPTER THREE
Research Design & Implementation
3.1 Major concerns in Fraud detection/prevention
3.2 The Proposed Methodology
3.2.1 Big Data platform
3.2.2 Deep learning model
3.2.3 Anomaly detection
3.3 Implementation of the Proposed system
3.3.1 Apache Spark Framework
3.3.2 H2O Work ow Framework
3.3.3 Data sets and design use
3.4 Summary
CHAPTER FOUR
Results & Analysis
4.1 Introduction
4.2 Performance measures
4.3 Experiment I - Results & Analysis
4.3.1 The Activation function
4.3.2 Between Deeper network & Epoch cycles
4.3.3 Multi-layer Feedforward Fraud predictor model
4.4 Experiment II { Results & Analysis
4.5 Summary
CHAPTER FIVE
5 Summary, Conclusions & Recommendations
5.1 Summary
5.2 Conclusions
5.3 Recommendations
Bibliography
Chapter 1
Introduction
1.1 Background to the Study
Fraud refers to the intentional illegal exploitation of a system which results in injury of an oblivious entity. Financial fraud involves the exploitation of nancial systems which results in the loss of nancial resource, the most prominent being monetary although other damages such as loss of integrity are possible. Fraud, waste, and abuse in many nancial systems are estimated to result in signi cant losses annually running into billions of US dollars.
Furthermore, the proliferation of the internet has exposed nancial systems to diverse fraudsters using di erent mechanisms to exploit nancial systems. This pro-vided an explode in attack patterns which rendered the once e ective case-based fraud detection solutionsno more e ective as the computational complexity increases with each new detected fraud. More seriously, their is a higher tendency for rst-time frauds going undetected. The case-based detection methods are also slow as a successful exploit could multiply if the solution took time to be integrated into the system. This problem can only be addressed with an online (on-the- y) adaptive (able to detect new frauds) solution.
Also of concern to nancial fraud detector solutions is, the prediction strength that indicates a Fraud detector's ability to correctly identify both known and novel Frauds. This is usually a direct function of how much fraud samples there are to model a solution. The emergence of Big Data and its Analytics has provided nancial fraud detection experts with verse amount of data that will enhance the detection models. Such solutions that use Big Data to model o er more comprehensive solutions.
A complete fraud detection model thus, must have the following properties:
1. Adaptive: This refers to the following abilities:
Ability to detect fraudulent activities within a short period of time. This is also referred to as its alertness.
Ability to detect rst-time fraudulent activities with high accuracy.
2. Predictive: This refers to the following abilities:
Ability to detect all new instances of fraudulent activities that have hap-pened in the past. This is very di cult to achieve if there is no data with
a considerable description of previous transactions.
Over the years solutions have been proposed to provide e ective solutions to - nancial frauds. Most of the models proposed to address the Fraud detection model property 1 have been statistical models that try to detect outliers in the data set (See [27], [21] and [8]). This follows after the assumption that fraudulent transactions will behave abnormally di erent from legitimate transactions. An abnormal pattern of behavior (i.e. an Outlier) is agged \suspicious." More recent, Machine Learning methods have been used to develop more e ective models (See [4], [11], [7] and [9]).
================================================================
Item Type: Project Material | Size: 44 pages | Chapters: 1-5
Format: MS Word | Delivery: Within 30Mins.
================================================================
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.