A Novel Method to Predict the Accuracy of Loan Sanction using KNN Algorithm and Compare with Decision Tree Algorithm

In this study, we aimed to evaluate the accuracy of loan sanction predictions using two different algorithms. We analyzed how well each algorithm performs by applying them to a sample size of 20, divided equally into two groups. Group 1 used the KNN algorithm with 10 samples, while Group 2 used the Decision Tree algorithm with another set of 10 samples.To ensure the reliability of our results, we applied an independent sample analysis with a G power value of 80% and an alpha level of 0.05. The results showed that the KNN algorithm achieved an accuracy of 70.83%, whereas the Decision Tree algorithm performed slightly better with an accuracy of 74.47%. The difference between the two algorithms was statistically significant, with a two-tailed significance value of 0.032 (p < 0.05).