Implementasi Algoritma Neural Network Berbasis Particle Swarm Optimazation untuk Penentuan Kredit

Abstract

Bad credit is one of the credit risks faced by financial and banking industry players. Bad credit occurs if in the long run, financial institutions or banks cannot withdraw credit loans within the allotted time. Bad credit has a bad impact for credit providers in the form of risk of loss. Of course this should not be allowed to drag on and must be resolved. However, to guarantee accuracy in determining creditworthiness, an accurate algorithm is needed. Neural Network is an algorithm that can map and classify data input and can be used for the prediction of several variables that are input into this algorithm. To be able to add accuracy in predicting the determinants of credit worthiness, this algorithm is optimized with Particle Swarm Optimazation. The results of the study in the form of confusion matrix prove that the accuracy of Neural Networks based on Particle Swarm Optimazation has an accuracy rate of 96.67% and AUC results of 0.965