Evaluation Clustering methods

There are many evaluation measures available like entropy, recall, precision, F-measure, silhouette co-efficient, purity, inverse purity for improving cluster's accuracy, efficiency and result.

1.　　Recall=A/(A+B), where A is the true positive, B is the false negative

         Pecision = A/(A+C), where C is the false positive

　　   F-measure=2*Precision*recall/ (precision+recall)

2.      Purity, Silhouette co-efficient: 

         

转载于:https://www.cnblogs.com/kellytran/p/5986941.html