Fraud Detection using Isolation Forest Algorithm in Credit Card



Author Name- Mrs. Vandana Tripathi(vandy3585shi@gmail.com)
Divyanshu Mishra(divyanshumishra45186@gmail.com)
Dhruv Gupta(dg739390@gmail.com)
Prakhar Tiwari(t.prakharofficial@gmail.com)
Department of Computer Science, Babu Banarasi Das Institute of Technology and Management, Lucknow, India

Abstract: Nowadays, everyone uses credit cards for a variety of transactions because of the demonetization of the economy. Therefore, there will be a higher likelihood of fraud. Banks maintain several, vast datasets. These data repositories can be used to extract crucial business data. Fraud is a problem having wide-ranging effects on the financial sector, the government, corporate sectors, and regular consumers. In recent years, the issue has arisen with an increase in reliance on new technologies like cloud and mobile computing. Physical detections are not only expensive and time-consuming, but they also don’t produce reliable results. It is hardly unexpected that financial and computationally based automated processes have been adopted by economic institutions. Traditional methods relied on labor-intensive manual methods like auditing, which arebecause of the complexity of the issue, inefficient and unreliable. The capacity of data mining-based systems to spot minute anomalies in huge data sets has demonstrated its value. So, in order to detect fraud, we have employed several supervised algorithms, which produce reliable findings. There are numerous sorts of fraud and various data mining techniques, and study is always being done to determine the optimal strategy for each situation. Although the phrase “financial fraud” has several potential connotations, for our purposes, it can be defined as the intentional use of illegal means or means of doing things in order to earn financial advantage.Fraud has a significant detrimental effect on society and industry; just credit card fraud alone results in billions of dollars in annual revenue loss. For the classification of authentic and fraudulent records from the dataset, we used isolation forests.

Keywords: Fraud detection, Decision tree, Isolation Forest Algorithm

I.INRODUCTION
Fraud is the misappropriation of a profit-making organization’s system without necessarily resulting in overt legal issues. The general act of deceiving another person or organisation in order to gain financial gain is fraud. Detecting credit card fraud involves classifying fraudulent transactions into two categories: legitimate transactions and hoaxes. The classic card-related frauds, internet frauds, and other types of fraud can all be generally grouped into three groups. Customer fraud and external fraud are terms used to describe fraud performed by people outside of the business, whereas management fraud and internal fraud are terms used to describe fraud committed by top-level management.Fraud detection automates and helps decrease the manual components of a screening



fraud-detection-in-credit-card
FREE DOWNLOAD

@ engpaper.com published paper
PUBLICATION PROCEDURE WITH US ENGPAPER.COM
ENGPAPER.COM PUBLISHED PAPERS