Hello, thanks for commenting. You can think of the problem as unsupervised learning. If you have some ground truth labels available then it heps. Else you can check the clustering performance evaluation page by Scikit-learn (https://scikit-learn.org/stable/modules/clustering.html#clustering-performance-evaluation). Also check the stack answer page for various other methods (https://stats.stackexchange.com/questions/79028/performance-metrics-to-evaluate-unsupervised-learning). Cheers!