I am working on clustering on binary data which has 25 features,

sample Feature 1 Feature 2 Feature 3 ...... Feature 25
1 1 0 0 011101 1
2 0 1 0 010011 0
3 1 0 1 101001 1

and I have used the Silhouette score to choose the number of clusters using the K-modes algorithm, but the score was very low. silhouette score using K-modes.data distribution using K-modes I have also tried the HDBSCAN algorithm using Jaccard and hamming distance metric. The silhouette score (around 0.26) was higher than the one using K-modes, but the data distribution was quite unbalanced. data distribution using HDBSCANTherefore, I would like to ask that are there other better clustering methods for binary data, more appropriate metrics to choose the number of cluster and evaluate the quality of clustering?


1 Answer 1


Look at every sample as a string and calculate any string similarity (one example is Hamming distance). After calculating all similarities, you will have the similarity matrix a.k.a Affinity Matrix. Then You are all set for Spectral Clustering. Comment here if you had any further questions.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.