Research Article
Improvement of Customer Loans Prediction Accuracy in Neural Networks
@INPROCEEDINGS{10.4108/eai.17-12-2022.2333265, author={Irwan Syah and Amron Amron and Herry Subagyo}, title={Improvement of Customer Loans Prediction Accuracy in Neural Networks}, proceedings={Proceedings of the International Conference on Economic, Management, Business and Accounting, ICEMBA 2022, 17 December 2022, Tanjungpinang, Riau Islands, Indonesia}, publisher={EAI}, proceedings_a={ICEMBA}, year={2023}, month={6}, keywords={neural network; svm-rfe; bagging; classification; optimaze selection}, doi={10.4108/eai.17-12-2022.2333265} }- Irwan Syah
Amron Amron
Herry Subagyo
Year: 2023
Improvement of Customer Loans Prediction Accuracy in Neural Networks
ICEMBA
EAI
DOI: 10.4108/eai.17-12-2022.2333265
Abstract
Credit screening is divided into two categories of credit applicants based on their potential repayment capacity. A suitable applicants will be accepted high probability of default on installment obligations. Predicting creditworthiness is one of the considerations in granting credit to customers. Therefore the prediction accuracy results are needed. The neural network algorithm integrated use bagging and feature selection Support Vector Machine Recursive Feature Elimination (SVM-RFE). This study showed that this method improved the accuracy of predictive model performances. The lowest accuracy value was 86.52%. While 87.15% using the neural network model, joint bagging and optimizing the SVM-RFE selection. The experiment results revealed that the neural network model with bagging and optimizing the SVM-RFE selection could improve the performance of at customer credit prediction model.