Information-based clustering.

@article{citeulike:445717, abstract = {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.}, address = {Joseph Henry Laboratories of Physics, and Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ 08544.}, author = {Slonim, Noam and Atwal, Gurinder Singh S. and Tkacik, Gasper and Bialek, William }, citeulike-article-id = {445717}, doi = {http://dx.doi.org/10.1073/pnas.0507432102}, issn = {0027-8424}, journal = {Proc Natl Acad Sci U S A}, keywords = {algorithm, clustering, information, theoretical}, month = {December}, posted-at = {2008-01-01 04:14:44}, priority = {2}, title = {Information-based clustering.}, url = {http://dx.doi.org/10.1073/pnas.0507432102}, year = {2005} }

TY - JOUR ID - citeulike:445717 L3 - citeulike-article-id:445717 TI - Information-based clustering. JF - Proc Natl Acad Sci U S A AD - Joseph Henry Laboratories of Physics, and Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ 08544. SN - 0027-8424 N2 - 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. KW - algorithm KW - clustering KW - information KW - theoretical AU - Slonim, Noam AU - Atwal, Gurinder Singh AU - Tkacik, Gasper AU - Bialek, William PY - 2005/12/13/ UR - http://dx.doi.org/10.1073/pnas.0507432102 ER -