Description
Abstract: Currently there are billions of users of cell phones and portable devices. With this growth of mobile technology, it comes with a lot of applications; one of them is applying for a soft quick loan. However, since financial institutions offering such services have limited resources and assets, it is required of them to check several factors before approving a mobile loan for a customer. This paper proposes a hybrid machine learning model for predicting whether a customer is credit worthy or not. This paper experiments; using existing financial institution data for both approved and disapproved customers for mobile loan facilities. The proposed model is a hybrid of machine learning techniques namely SVM, regression, KNN, and decision trees. Our model is further trained and tested on the various datasets. Machine learning and data mining experiments are conducted using existing ML libraries. The results and discussions are presented in the form of descriptive statistics and prediction metrics.
Keywords: Loan, Machine Learning, mobile, model, Training, Testing, Prediction.
