Machine Learning Models Based Data Quality Analysis to Detect Credit Card Frauds

Authors

Amit Pundir, M.Tech Scholar, Rajesh Pandey, Assistant Professor
Shobhit Institute of Engineering & Technology, Meerut, Uttar Pradesh, India.

Abstract

Considering the broader consequences and taking into account the multiple cycles, cash distortion is a serious issue in money-related business. The management of quality data with mining of data has been successfully applied to datasets to automate the examination of vast measures of complex data. Similarly, data mining has played a major role in isolating frauds like credit cards during online trades. Detection of fraud at credit card is the management of quality data issue considered under data mining, attempted for two important reasons – first, the profiles of standard and fraudulent practices change regularly, and furthermore, clarifications are needed. Second, credit card, extortion data is surprisingly sluggish. This research paper examines the performance of Decision Trees, Logistics Regression, and Random Forests that rely on deliberately heavily tilted credit card fraud data. The dataset of credit card transactions is sourced from Kaggle (a publicly available dataset repository) with a total number of transactions of 284,807. These strategies are applied to values of raw data and data processing methods. Evaluation of the performance of methods depends on accuracy, sensitivity, specificity, precision, and recall. The results show ideal accuracy for decision trees, logistics regression, and random forest classifiers with 90.8%, 98.5%, and 99.1% respectively.