Hybrid K-Means Clustering for Training Special Children using Utility Pattern Mining

Abstract

In this Hybrid K-Means Clustering for Grouping research work, the k-means methodology is a well-known process for grouping things together. Most of the time, this algorithm sorts the objects into a set number of clusters, but in this case, the user gives the number k. At first, it picks cluster centres at random and measures how far apart k points are. This kind of cluster centre is called k centroids, and it will keep changing until there are no more changes. When making applications that use machine intelligence, a machine should be able to think like a person and make the right choices. In this case, it's not possible to get k-points from the user. So, the Genetic Algorithm (GA) is used to search with heuristics to find the initial cluster centres. The goal of this research work is to look at how k-means clustering with GA can be used to optimise. The performance evaluation of hybrid k means method illustrates the precision and accuracy of the selected clustering methods. The result states that precision was 78.35% and its accuracy was 72.67% found while using the approach. The ROC curve analysis is performed with sensitivity and specificity data. The result states that the area under curve of the approach is 84.0%.