KOMPARASI ALGORITMA KLASIFIKASI DATA MINING DALAM PENENTUAN RESIKO KREDIT PADA KOPERASI SERBA USAHA

Abstract

Growth and development of the cooperative lately , where lending credit cooperatives that
cater to its members to develop business in the workshop , stores , services and so on . No doubt ,
lend funds to member cooperatives will surely emerge problems , such as members of the
borrower paying the overdue installment of funds , misuse of funds for other purposes , the
customer fails to develop its business so as to result in cooperative funds do not flow or it can lead
to bad credit . In this study, using the method of comparison between the three models , namely
Naive Bayes , Neural Network and k - nearest neighbor , the cooperative member data to
determine which is the most accurate method for determining credit risk in business cooperatives
that will produce cooperative members who pay installments smoothly or delinquent in the
payment of the loan installments . From the results of the study to measure the performance of the
three algorithms using the test method validation . Matrikx confusion , note that the k - nearest
neighbor method has the highest accuracy value of 93.00 % , for the Naive Bayes method has an
accuracy value of 90.33 % , while the method which has the lowest value is the Neural Network
method has an accuracy value of 85.67 % . As for the value of Area Under the Curve ( AUC ) for
the k - nearest neighbor method has the highest score is 0989 , Naive Bayes method is 0.619 ,
while the lowest value method with the method of Neural Network 0.467 .
Keywords : comparison , naive Bayes , neural network , k - nearest neighbor .