<![CDATA[Providing credit is one of the bank offers offered to customers, but extending credit to customers who are not appropriate can cause problems such as customers who do not pay installments on time and even delay payment of installments for several months until bad credit occurs so that this can be detrimental to the bank. Therefore, in this study a comparative method will be carried out to find out which method is the best in classifying the smoothness of bank credit payments. It is hoped that the results of the research can be used as material for consideration by the bank in the selection of bank credit customers. In this study using a dataset from the UCI Machine Learning Repository, the credit payment data totaled 29,998. The dataset is split by dividing 70% train data and 30% test data with the amount of each data, namely 24000 train data and 6000 test data. Meanwhile, the labels used are Eligible and Ineligible. In this study, implementing the data mining process using the CRISP-DM framework and using the Python programming language. From the results of the evaluation using the confusion matrix, the best accuracy value was obtained for the random forest algorithm, namely 82.22%, precision of 80.44%, recall of 82.22% and f1-score of 80.0%. Meanwhile, the KNN algorithm obtains an accuracy value of 81.55%, a precision of 79.5%, a recall of 81.55% and an f1-score of 79.11%. Based on the results of this evaluation, the Random Forest algorithm has the best accuracy compared to the KNN algorithm in classifying bank credit payments.]]>
Copyrights © 2024
