<![CDATA[The purpose of this study is to demonstrate that smartphone usage data can be used to make predictions and find the best classification method for Credit Scoring even if the sample dataset is small. The study used a classification algorithm to divide customers into paying and arrears customers using mobile data, and then compared the predicted results with the actual results. There are several related works that are publicly accessible where mobile data has been used for Credit Scoring, but all of them use much larger data samples. Small companies cannot use large data sets as used in previous studies. In this paper the author tries to argue that there is data smartphone has good predictive power even though the dataset is small. The author concludes that with a sample data consisting of smartphone data as many as 5,702 debtors can still predict credit risk well. The best classification method using random forest yielded a result of 0.68 AUC equivalent to a ratio of 0.36Keywords : credit scoring, classification, smartphone data]]>
Copyrights © 2023
