Improvement of personal loans granting methods in banks using machine learning methods and approaches in Palestine

For banking organizations, loan approval and risk assessment which is related is a very complex and significant process which is needs a high effort for relevant employee or manager to take a decision, because of manual or traditional methods that used in banks. The banking industry still needs a more precise method of predictive modeling for several problems. In general, for financial institutions and especially for banks forecasting credit defaulters is a hard challenge. The primary role of the current systems is to accept, or sending loan application to a specific level of approval to be studied and it is very difficult to foresee the probability of the borrower for paying the due dues amount without using methods to predict. Machine learning (ML) techniques and the algorithm that belongs to are a very amazing and promising technique in predicting for a large amount of data. Our research proposed to study three machine learning algorithms [1], Decision Tree (DT), Logistic Regression (LR), and Random Forest (RF), by using real data collected from Quds Bank with a variables that cover credit restriction and regulator instructions. The algorithm has been implemented to predict the loan approval of customers and the output tested in terms of the predicted accuracy.