2

I have selected 'Nursery' data set from UCI machine learning repository and run 2 different clustering algorithm on, K Means and Hierarchical clustering. How should I compare these to algorithm to see which one did a better job on the data set?
I have calculated the parameters:

  • adjust_rand_score
  • adjust_mutual_info_score
  • silhouette_score

for each 2 clustering model. Is that true to compare the calculated parameters to see which one did better?
For example: K Means model:

  • adjust_mutual_info_score = 0.044
  • silhouette_score = 0.56

For Hierarchical Clustering model:

  • adjust_mutual_info_score = 0.16
  • silhouette_score = 0.59

So can I say the Hierarchical model has did a better job? Or there is another parameter to compare these 2 clustering algorithms?

  • I wonder, is there no answer to your question already? [Comparing clustering results](https://stats.stackexchange.com/search?q=compare+clustering) seems like frequently asked thing... – Jan Kukacka Jun 28 '18 at 09:44

0 Answers0