Information-based clustering

Noam Slonim, Gurinder Singh Atwal, Gašper Tkačik, William Bialek

Research output: Contribution to journalArticlepeer-review

190 Scopus citations


In an age of increasingly large data sets, investigators in many different disciplines have turned to clustering as a tool for data analysis and exploration. Existing clustering methods, however, typically depend on several nontrivial assumptions about the structure of data. Here, we reformulate the clustering problem from an information theoretic perspective that avoids many of these assumptions. In particular, our formulation obviates the need for defining a cluster "prototype," does not require an a priori similarity metric, is invariant to changes in the representation of the data, and naturally captures nonlinear relations. We apply this approach to different domains and find that it consistently produces clusters that are more coherent than those extracted by existing algorithms. Finally, our approach provides a way of clustering based on collective notions of similarity rather than the traditional pairwise measures.

Original languageEnglish (US)
Pages (from-to)18297-18302
Number of pages6
JournalProceedings of the National Academy of Sciences of the United States of America
Issue number51
StatePublished - Dec 20 2005

All Science Journal Classification (ASJC) codes

  • General


  • Cluster analysis
  • Gene expression
  • Information theory
  • Rate distortion


Dive into the research topics of 'Information-based clustering'. Together they form a unique fingerprint.

Cite this