Fraud Detection Through the Utilization of Machine Learning
Grade Level at Time of Presentation
Secondary School
Major
Master of Business Administration
Institution
Western Kentucky University
KY House District #
20
KY Senate District #
32
Faculty Advisor/ Mentor
Dr. Lily Popova Zhuhadar
Department
Information Systems Department
Abstract
The Association of Certified Fraud Examiners (ACFE) reported [1] that a typical organization loses approximately 5 percent of its total revenue in a given year as a result of fraud, with an average loss per case of $2.7 million. Financial fraud and identity theft are two significant risks all consumers face when they use an electronic transaction to make a purchase. Credit card companies rely on sophisticated software to ensure its customers are protected from financial ruin, but this software is not accurate all the time.
This research investigates irregularities in historical data to prove that the enterprise is not maintaining its accounting system in agreement with its procedures. Within this context, I plan to present a research study that utilizes machine learning algorithms to detect fraud in transactional datasets. The research process starts with the theoretical consideration known as Benford’s law (the law of anomalous numbers.) Benford’s law assumes that there is a relationship between the frequency of a particular digit and its magnitude. By using Benford’s law, I identify whether or not a customer (or a sales manager) exhibits a behavior (overtime) that corresponds with the normal pattern of behavior. To this end, I acquired a dataset from SAS Institution. This dataset consists of 101,822 anonymous transactional records. The developed machine learning algorithm detects deviated values among transactions from Benford’s law distribution. Finally, I examine how new strategies can be utilized to detect financial fraud in the lives of everyday people better. Besides, I present a framework on how to make sure human intervention is rarely needed for the process to work.
Fraud Detection Through the Utilization of Machine Learning
The Association of Certified Fraud Examiners (ACFE) reported [1] that a typical organization loses approximately 5 percent of its total revenue in a given year as a result of fraud, with an average loss per case of $2.7 million. Financial fraud and identity theft are two significant risks all consumers face when they use an electronic transaction to make a purchase. Credit card companies rely on sophisticated software to ensure its customers are protected from financial ruin, but this software is not accurate all the time.
This research investigates irregularities in historical data to prove that the enterprise is not maintaining its accounting system in agreement with its procedures. Within this context, I plan to present a research study that utilizes machine learning algorithms to detect fraud in transactional datasets. The research process starts with the theoretical consideration known as Benford’s law (the law of anomalous numbers.) Benford’s law assumes that there is a relationship between the frequency of a particular digit and its magnitude. By using Benford’s law, I identify whether or not a customer (or a sales manager) exhibits a behavior (overtime) that corresponds with the normal pattern of behavior. To this end, I acquired a dataset from SAS Institution. This dataset consists of 101,822 anonymous transactional records. The developed machine learning algorithm detects deviated values among transactions from Benford’s law distribution. Finally, I examine how new strategies can be utilized to detect financial fraud in the lives of everyday people better. Besides, I present a framework on how to make sure human intervention is rarely needed for the process to work.