how do I evaluate K-means clustering anomaly detection method as there is no labelled data of anomaly class. To find the cluster (K), I have used the silhouette score from Scikit learn library. Scikit learn K-means were used too.