Hybrid Cuckoo Search-GA based SVM Classifier Algorithm for Loan Approval Prediction

Abstract

            Banking has become an unavoidable activity and a necessary routine in our daily lives in order to carry out all types of financial dealings and transactions. The banks have a lot of items to sell because of profit and earrings, but the interest paid by the loaners is one of the most essential sources of income. To a large extent, the interest that a bank extracts from the loan amount is a crucial criterion for predicting the profit of a bank. As a result, it's critical for a financial system to keep the loan approval process running smoothly by analyzing and learning from diverse criteria. Maintaining loan defaulters and non-functioning assets is also unavoidable. By evaluating their general details and personal traits, a prominent machine learning model will be required to predict and award an accurate loan approval method. To analyze defaulters and measure their credit history and civil score, the proposed model (Cuckoo Search – GA based SVM Classifier) is applied. As a result, this model is critical in granting loans and identifying defaulters in order to maximize credit score and profit. In this research, machine learning methods are used to predict client loan approval: Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), Nave Bayes (NB), K-Nearest Neighbor (KNN), and Random Forest (RF). The experimental findings show that the suggested technique outperforms other machine learning algorithms in terms of accuracy.