I have a large number of string and I calculated their pairwise Levenshtein distance to form a distance matrix. I tried hierarchy clustring in Matlab but the results do not make much sense. I think it is because I do not use a built in distance metric. After formign linkage, the cophenetic correlation coefficient is pretty small or even negative. And I find difficult to write a customed distance function handle. Is there any other recommendation for clustering such a distance matrix? Or how could I write a customed distance function hanfdle?
Asked
Active
Viewed 6 times
