Hybrid Genetic Algorithm with a New Fitness Measure of Clusters for Training Special Children

Clustering is one of the most interesting and popular areas of data mining, which is gaining more and more interest from researchers. There has already been a lot of research, and as a result, many clustering algorithms have been made. K-means is the most common, simple, and well-known way to group numerical data into groups. But it has been changed over time so that it can also handle categorical data. People have noticed that there haven’t been many works done on categorical data clustering so far. This is likely because of how hard it is to work with them. When genetic algorithms are used in the field of clustering, they improve the quality of the clustering results, making them almost as good as the global best instead of just the local best. Taking these things into account, we chose chapter 4’s work area to include categorical data clustering and genetic algorithms. As the results, implemented the approach which described using the similarity measure and a consistent manner of encoding an item and a cluster representative as a fuzzy set and a new fitness measure has been developed that finds the most fit or fittest chromosome among all generations by measuring the fitness of each chromosome (clustering solution) belonging to a generation of population.