The clustering by k-means of using the target variable. To determine the number of clusters with the variance of the target variable in the cluster.