<![CDATA[To see whether the proposed research model is able to improve the performance of the classification of the Glass Type Identification data using the K-Nearest Neighbor (K-NN) method then the results will be compared with the C4.5 method and the Naïve Bayes method, a performance analysis of the methods will be carried out. The results are based on the results of the Confusion Matrix tabulation (two-class prediction. In this study, only three preprocessing processes were carried out. The first process is handling missing value. The missing value for attributes with numeric values is replaced by the mean (mean) value of the attributes in the same column. Meanwhile, the missing values for attributes with nominal values are replaced by the most likely values for the attributes in the same column. Then the second process is the handling of duplicated data. The data recorded were 214 data, the number of attributes was 9 attributes and the number of classes was 6 classes.The results of this study show that the highest accuracy value is in the C4.5 method with an accuracy of 73.45% with a value of K = 2 and an error rate of 26.55%, while the method with low accuracy is the KNN method. with an accuracy value of 61.95% and an error rate of 38.05%. Naïve Bayes has an accuracy of 63.33% and an error rate of 36.67. Therefore C4.5 is more effective than the two methods. ]]>
Copyrights © 2021
