ระบบสนับสนุนการตัดสินใจสำหรับการจำแนกระดับลูกค้าเงินกู้ของธนาคารเพื่อการเกษตรและสหกรณ์การเกษตรสาขาท่าวังผา โดยใช้ต้นไม้ตัดสินใจ

Abstract

The objective of independent study was to developed decision support system for debtor-client level classification of bank for agriculture and agricultural co-operatives Thawangpha Branch using decision tree. The result of this study will make it easy to work and reduce errors in the work. This system used data mining technique to create a model for debtor-client level classification in the form of a decision tree. This study used the collection of 1,500 instances, which were 1,000 instances of training set and 500 instances of testing set. The model formulation was perform based on Weka. The two types of decision tree were studied, i.e., ID3 and C4.5(J48) algorithm. The accuracy of these models was analyzed from confusion matrix. From the experiment, ID3 algorithm returned a higher accuracy than C4.5(J48) algorithm. The average of prediction accuracy was about 92.2%. Thus, ID3 algorithm was selected for implementing a classification model for debtor-client level. Then, the stand-alone application based on the derived model from ID3 was developed by using Netbeans with Java.