Naïve Bayes Algorithm Based on Forward and Backward Selection for Predicting Mortgage Loan Application Status

Andriano, Andriano
Harison, Harison

Electronic Journal of Education, Social Economics and Technology
Vol 6, No 1 (2025)

This study discusses about the classification analysis of the status of home ownership credit at one of the banks in Pekanbaru City using the naïve Bayes algorithm. The purpose of this study is to compare the results of naïve Bayes classification with variable selection using forward and backward selection without variable selection. Variables from the selection results that are included in the best model with transform data use Weight of Evidence method are the type of customer, residential status, last education, and occupation. The results show that the naïve bayes algorithm with variable selection is better than without variable selection. This is shown from the value of the confusion matrix, namely accuracy and specificity, which reached 57,10% and 72,32%, although the sensitivity value was relatively small, 33,10%.