I have a categorical dataset (survey data on a likert-scale: never to extremely often). So, I need to cluster them. Therefore, I am planning to use K-modes. But, I am stuck at selecting the optimal number of clusters. I noticed that Elbow Method, Silhouette, and sqrt (n/2) are in use to determine the number of clusters for K-means. But could not find anything proper or K-modes. I am using Python and also not a clustering expert.