Applicability of K-Means and Genetic Algorithm in Clustering of Indian Mustard Genotypes

The purpose of the study was to check the efficacy of K-means and Genetic Algorithm methods for clustering of Indian mustard genotypes. The secondary data for the growth and yield characteristics of 80 Indian Mustard genotypes were used to identify patterns and best genotypes for plant breeders through K-means and Genetic algorithm clustering methods. The maximum RV-coefficient criterion was used to find best subset size of 3 variables. The clustering was done using K-means and Genetic Algorithm methods based on subset size of 3 variables and all 12 variables. The applicability of both methods was compared from obtained results. It was concluded that the quality of clusters based on the percentage of between sum of squares (BSS) was best in case of K-means method using a subset of the variables.