PENGAPLIKASIAN ALGORITMA CLASSIFICATION BASED ON ASSOCIATION UNTUK KLASIFIKASI RESIKO PEMBERIAN KREDIT (Studi Kasus: PT. Telkom CDC Sub Area Kupang); APPLICATION OF CLASSIFICATION BASED ON ASSOCIATION ALGORITHM FOR CREDIT RISK CLASSIFICATION (Case Study: PT. Telkom CDC Kupang Sub Area)

Robynson Willson Oktovianus Amseke, Edi Winarko

2013 | Disertasi | PROGRAM STUDI S2 ILMU KOMPUTER

One of the causes of non-performing loans come from the internal, is caused by a lack of rigorous team in conducting the survey and analysis, or it could be due to subjective evaluation and analysis. This can be solved using other selection techniques that have the thoroughness of analysis and objective assessment, the computer application that uses data mining techniques. Data mining technique was used in this study to classify credit risk by applying algorithms Classification Based on Association (CBA). This algorithm is an algorithm classification of data mining which integrating association and classification techniques. Preprocessed initial-credit data, will be processed using the CBA algorithm to create a model of which is to classify the new loan data into swift class or bad one. Testing techniques the accuracy of the model was measured by 10-fold cross validation. The result shows that the accuracy average value using the CBA algorithm (57,86%), was slightly higher than those using the algorithms of SVM and Naive Bayes from Rapid Miner 5.3 software (56,35% and 55,03%, respectively).

Kata Kunci : classification based on association; CBA; data mining; klasifikasi; resiko pemberian kredit