COMPARATIVE ANALYSIS OF K-NEAREST NEIGHBOR AND NAÏVE BAYES CLASSIFIER ALGORITHMS FOR THE FEASIBILITY CLASSIFICATION OF LIEN RECEIPT

Abstract

One of the problems faced by companies engaged in pawn services is that they must always pay attention to the appropriateness of the value of the loan based on the goods being pawned. Sometimes the pawnbrokers do not come back after depositing their goods with the company and do not pay them off. If the value of the goods does not match and the amount of money lent is too large compared to the value of the pawned goods, the company will lose money. There needs to be a way to determine whether or not it is appropriate for a pawned item to be accepted by the company and the value of the loan given to the pawnbroker. The way to find out can be classified by data mining classification techniques using the K-Nearest Neighbor method and the Naïve Bayes method. The analysis carried out by manual calculations and testing with Rapidminer tools resulted in the accuracy of the comparison of the two methods of the K-Nearest Neighbor and Naïve Bayes algorithms in predicting the classification of the feasibility of receiving pawned goods using the Confusion Matrix formula which resulted in an accuracy rate of the K-Nearest Neighbor method of 40%. while the Naïve Bayes method has an Accuracy Value of 20%, so we can determine that the accuracy of the prediction value using the K-Nearest Neighbor method is better than the Naïve Bayes method.