title: From Instance-level Constraints to Space-Level Constraints: Making the Most of Prior Knowledge in Data Clustering
creator: Klein, Dan
creator: Kamvar, Sepandar D.
creator: Manning, Christopher D.
subject: Computer Science
subject: Data Mining
subject: Miscellaneous
description: We present an improved method for clustering in the presence of very limited supervisory information, given as pairwise instance constraints.  By allowing instance-level constraints to have space-level inductive implications, we are able to successfully incorporate constraints for a wide range of data set types.  Our method greatly improves on the previously studied constrained k-means algorithm, generally requiring less than half as many constraints to achieve a given accuracy on a range of real-world data, while also being more robust when over-constrained.  We additionally discuss an active learning algorithm which increases the value of constraints even further.
publisher: Stanford
date: 2002-02
type: Techreport
type: NonPeerReviewed
format: application/pdf
identifier: http://ilpubs.stanford.edu:8090/528/1/2002-10.pdf
identifier: Klein, Dan and Kamvar, Sepandar D. and Manning, Christopher D. (2002) From Instance-level Constraints to Space-Level Constraints: Making the Most of Prior Knowledge in Data Clustering. Technical Report. Stanford.
relation: http://ilpubs.stanford.edu:8090/528/