Comparison of the K-Means Algorithm and C4.5 Against Sales Data

Abstract

In general, the process of collecting and grouping data requires a long process. And if it has to be grouped manually it takes a very long time. Therefore, data mining is a solution for clustering data - a lot of data to classify it. In this research conducted at CV.Togu - Togu On Medan Branch, data mining is applied using the K-Means process model and the C4.5 algorithm which provides a standard process for using data mining in various fields used in classification because the results of this method easy to understand and easy to interpret. . The K-means method is a non-herarical method which is an algorithmic technique for grouping items into k clusters by minimizing the distance of the SS (sum of square) to the cluster centroid. In the K-means method, the number of clusters can be determined by the researcher himself. And the testing methods used to measure cluster quality are the Silhouette Coefficient and the Elbow Method. Based on the research conducted, there are significant differences before and after using the two methods. The results of the K-Means algorithm will be compared with the results of the C4.5 algorithm in the form of rules (decision trees). This research produces data on goods that have the highest level of sales/behavior