IMPLEMENTATION OF MACHINE LEARNING MODEL IN DETECTING CREDIT CARD FRAUD

ABSTRACT

Credit card fraud includes both physical credit card loss and the loss of confidential credit card information. Numerous machine learning algorithms can be used for detection. This study presents three algorithms for identifying whether transactions are genuine or fraudulent. Data on credit card fraud detection were utilised in this study. For the corporations who issue credit cards, it is crucial to be able to prevent clients from being charged for items they did not buy, by identifying fraudulent credit card transactions. Such problems can be solved using machine learning. This project demonstrates how to model datasets using machine learning and credit card fraud detection. Modeling previous credit card transactions based on data that has been shown to be duplicitous is the problem with credit card fraud detection. Then, a new transaction's legitimacy is determined in light of this pattern. The goal is to detect 100%, the amount of fraudulent transactions while classification of theft. Credit card fraud detection is a typical example of classification. In this process, I focused on the analysis, preprocessing of datasets and the application of several outlier detection algorithms such as Naive Bayes, Logistic Regression and Random Forest to PCA transformed credit card transactions. The findings demonstrate that each algorithm is highly accurate at detecting credit card fraud.