Loan Credibility Prediction using Naïve Bayes, Random Forest, and Logistic Regression: A Comparative Study

Data mining is a rapidly developing business application field. All banks consider the loan approval process as one of the most significant processes. This research paper conducts a comparative study between three major classifiers, namely: Naïve Bayes, Random Forest, and Logistic Regression in the field of loan approval and credibility. The methodology used in this research is CRISP-DM. This research compares the three classifiers in terms of efficiency to see which one can be used to predict whether a loan applicant would be accepted or denied to support the decision making process. The highest classifier approach in terms of accuracy was Logistic Regression, which had an accuracy rate of 80.1%, as compared to Naïve Bayes and Random Forest that both showed accuracy of 78.9%. Other measures were taken to back up the conclusion using Weka as an analysis tool to support the results and findings.