PERBANDINGAN ANALISIS DISKRIMINAN FISHER DAN NAIVE BAYES UNTUK KLASIFIKASI RISIKO KREDIT (Studi Kasus Debitur di Koperasi Jateng Amanah Mandiri Cabang Sukorejo Kendal)

Abdur Rofiq (Universitas Diponegoro)
Triastuti Wuryandari (Universitas Diponegoro)
Rita Rahmawati (Universitas Diponegoro)

Jurnal Gaussian
Vol 5, No 1 (2016): Jurnal Gaussian

Credit is a form of money lending to debitors conducted by financial institutions such as cooperatives. In practice there are obstacles in the form of bad credit. Analyze by Fisher discriminant analysis method and Naive Bayes is used to classify the debitors fall into the category bad debitorr or not. This study uses data from  the Debitors of Cooperative of Central Java Amanah Independent in Sukorejo Kendal Branch. The data obtained is used for classification by Fisher discriminant analysis and Naive Bayes method. Data obtained has  multivariate normal distribution, has the same of variance-covariance matrix and has metric scale. Fisher discriminant analysis and Naive Bayes calculated and compared to the level of accuracy. From this research, the degree of accuracy of each method, namely 90% for Fisher Discriminant Analysis and 83.33% for the Naive Bayes. Having tested using the proportion test, Fisher discriminant analysis method is no different accuracy when compared with Naive Bayes to classify credit risk. Keywords: debitors, credit risk, Fisher discriminant analysis, Naive Bayes.