Classification Of Credit Customers of Mutiara Sharia Savings, Loans and Financing Coorporative Using The KNN Method

Abstract

This study applies the K-Nearest Neighbor (KNN) algorithm to classify the credit status of savings and loan cooperative members into two categories, namely current credit and non-performing credit, based on the variables provided. The data source for this study is data collected by the researcher directly from the primary source, namely the Muharjam Kolaka savings and loan cooperative. Several data collection techniques include: observation methods, interviews, and literature research. The classification process begins with a data preprocessing stage including handling missing data (missing values), data normalization and transforming categorical data into numerical. Overall, the results of the study indicate that the KNN algorithm has good potential as a decision-making tool in the creditworthiness analysis process in cooperatives. This study divided the training data and testing data by 20% or 0.2 where the total number of data used was 614, the testing data became 123, the training data 491 after the 2% division. Next, determine the K value, the K value used was K = 11. In the performance testing phase of the model developed, this study used a confusion matrix, with an F1-score of 90%, a recall of 98%, and a precision of 83%. The accuracy was 0.83, or rounded up to 0.84, or 84%. Both computerized and manual calculations yielded similar results according to the accuracy formula in the confusion matrix.