Application of X-Means Method for Grouping Early Childhood Diseases

Abstract

Grouping can use clustering to group data based on the similarity between the data, so that the data with the closest resemblance is in one cluster while the different data is in another group. The X-Means algorithm is the development of K-Means. The weakness of X-Means is that in determining the distance matrix, the distance matrix is ​​an important factor that depends on the X-Means algorithm data set. The resulting distance matrix value will affect the performance of the algorithm. The results of the study are: testing with variations in the number of centroids (K) with values ​​of 2,3,4,5,6,7,8,9,10. The author concludes that the number of centroids 3 and 4 has a better iteration value compared to the number of centroids that are getting higher and lower based on the iris dataset with the jarax matrix Manhattan Distance. From the test results with the X-Means cluster point, calculate the Euclidean Distance distance with 100 iris data reaching the 9th iteration, while with 100 iris data by calculating the Manhattan Distance distance it reaches the 10th iteration. Meanwhile, in determining the cluster point using the X-Means method from 100 data iris reaches its 7th iteration.