Artificial Intelligence Based System for Bank Loan Fraud Prediction

Financial institutions need advanced, current, and customized predictive analytics to protect themselves from the frustrating fraudster. Artificial Intelligence, Machine learning and statistical methods are in high demand from data scientists and statisticians who understand them; thus the demand for them is growing recently. The alarming rate by which loan beneficiaries default banks have course a lot of losses among many banks, and deprived many potential beneficiaries of access to the loan. This fallacy leads to inefficient and/or inaccurate management of loans in banks, and sadly, many banks have close down and have not yet realized that the labour-intensive approaches to loan management are not efficient enough. The trend has caused many banks workers to lose their job. The traditional ways of detecting fraud in bank loan management are not effective because the credit officer can easily be manipulated and not even discovered many loan defaulters. Therefore, this paper used Artificial Neural Network to detect loan fraud in bank loan management to avoid loan defaulter manipulate the officer in charge of loan administration. A loan credit dataset of 600 customers in a microfinance bank was used in this study. However, relevant features from the dataset were extracted to build a model that yields 98% accuracy. This promising solution makes fraud detection easier, and as well provide support to the bank to detect fraud in loan management.