Bank fraud detection for know your customer (KYC) in fintech

Ong, Belynda Liyen (2026) Bank fraud detection for know your customer (KYC) in fintech. Final Year Project, UTAR.

PDF Download (12Mb)

Abstract

With the rapid development of FinTech, the integration of ML into KYC processes has emerged as a solution to enhance bank fraud detection. This research explores the application of ML techniques to enhance fraud detection in banks within the context of KYC procedures in the FinTech industry using a dataset from the open-source platform Kaggle to test ML algorithms. EDA and feature engineering using PCA were also carried out. The ML techniques highlighted in the research are supervised learning techniques, namely LR, SVM, DT, RF, and XGBoost, which were used to build a robust model for detecting fraudulent transactions in banks. The model's performance is evaluated through the Confusion Matrix, which provides a comprehensive assessment of classification accuracy, precision, recall, and F-Measure to show the effectiveness of fraud detection. The models are trained and tested on financial transaction data to detect suspicious patterns that indicate fraud. The platforms used to conduct the experiments were Jupyter Notebook and Kaggle, using the Python programming language to conduct data collection, data preparation, data preprocessing, feature engineering, modelling, and performance evaluation to ensure effective and efficient algorithm implementation. Finally, the best-performing model is integrated into the visualisation application Power BI to provide real-time monitoring of detected fraudulent transactions and visualise trends in potentially fraudulent activity. This research contributes to the development of a fraud detection system that aids behavioural KYC regulatory compliance in financial institutions and contributes to the body of knowledge by encouraging further experimentation and improvements in integrating ML techniques alongside KYC regulatory compliance.

Actions (login required)