calm hedge
Perhaps someone can help me with a statistics-related data analysis issue. I'm having a pretty hard time understanding the following, translated from the attached image:
Clustering quality assessment: Silhouette index
The silhouette index is an effective indicator for validating a clustering method. For any individual i in the data set, the silhouette index is defined by the following formula: s(i) = (b˅i - ai) / max(a˅i, b˅i) where a˅i is the average dissimilarity between individual i and all the other individuals of the group G˅j to which he belongs and b˅i is the minimum of the average dissimilarities between the individual i and all the other individuals of the groups G˅k (k = 1, 2, ..., c with k <> j).
rich oasis