FRAUDULENT TRANSACTION IN THE FIELD OF FINANCE

ABSTRACT Financial scam is a crucial problem in the banking industry and the detection of fraudulent transactions is an important task for banks to protect their customers and maintain confidence in the financial system. This has led to an exponential rise in daily transactions. In this article, three to four methods are used to form normal transactions and fraudulent transactions, namely (Support Vector Machine), Logistic Regression, Decision Tree, Random. Forest, Naïve Bayes, ANN, Bagging, Boosting and K-Nearest Neighbour. The tools and technologies used are sklearn, imblearn and benchmark metrics to evaluate model performance. In addition, in this paper, we want to combine the model that gives the most accurate accuracy result, using the classification ratio to check the accuracy and recall value of the model and help detect the fraudulent transaction faster. This evaluation provides comprehensive guidance for choosing an optimal algorithm based on the type of fraud, and this article shows the result with an appropriate performance metric. Keywords: Financial fraud, ML technique