Fraud Detection of Credit Card Using Logistic Regression

Abstract

In modern times, credit card theft has developed a major concern for banks, as identifying fraud in the credit card system has become increasingly difficult. Machine learning plays a key role in detecting credit card fraud in transactions in order to address this challenge. Banks utilize a variety of machine learning techniques to forecast these transactions, collecting historical data and adding new variables to improve prediction capabilities. The suggested method builds the classifier using logistic regression to avoid credit card fraud. A pre-processing phase is employed to handle dirty data and ensure high detection accuracy. To clean the data, the preprocessing step employs two innovative essential strategies: the mean-based technique as well as the clustering-based technique. They are frequently confused with valid approaches that compare both fraud and normal data, but this is never enough to detect fraud adequately. This research shows how machine learning can be used to detect credit card fraud. Credit Card Fraud Detection is a project that shows how to use machine learning to model a knowledge set. Credit card transaction modelling, which has previously been done with fraud transaction data, is part of the Credit Card Fraud Detection problem. Our system will examine a new transaction to see if it is fraudulent. Our aim is to detect all suspicious transactions with as few false positives as possible. A transaction's likelihood of being fraudulent will be determined by our algorithm.

Keywords: Logistic regression, Sigmoid curve, seaborn, classification

JEL Classification: C25, C24