High Risk Bank Loan Recognition Based on Machine Learning

DOI

10.2991/aebmr.k.220405.357How to use a DOI?

Keywords

Bank Loan; Machine Learning; Random Forest Model; Decision Tree Model

Abstract

This paper aims to compare the performance of C5.0 decision tree and random forest in high-risk bank loans’ recognition. The author mine 1000 loan information data containing 16 variables and find our model’s accuracy is around 0.71. Then the author try to do some change on model parameters to improve the model. The author finds that increasing the cost of false negative in error matrix can help bank better avoid the high risk loan. Compared different trials in decision tree, the author find that the model perform the best when trials equal to 45. Compared different mtry in random forest, the author find that the model performs the best when the mtry equal to 11. Then the author compares the random forest and decision tree according to ROC, sensitivity and specificity. The results show that random forest has strength in ROC and sensitivity, while decision tree has strength in specificity. And due to the random forest’s bigger AUC value, the author conclude that random forest model slightly outperformed the tree model.

Copyright

© 2022 The Authors. Published by Atlantis Press International B.V.

Open Access

This is an open access article distributed under the CC BY-NC 4.0 license.

Download article (PDF)

TY  - CONF
AU  - Zishen Zhao
PY  - 2022
DA  - 2022/04/29
TI  - High Risk Bank Loan Recognition Based on Machine Learning
BT  - Proceedings of the 2022 7th International Conference on Social Sciences and Economic Development (ICSSED 2022)
PB  - Atlantis Press
SP  - 2118
EP  - 2121
SN  - 2352-5428
UR  - https://doi.org/10.2991/aebmr.k.220405.357
DO  - 10.2991/aebmr.k.220405.357
ID  - Zhao2022
ER  -