Department of Statistics Unitmark
Dietrich College of Humanities and Social Sciences

Clustering with Confidence: A Binning Approach

Publication Date

November, 2008

Publication Type

Tech Report

Author(s)

Rebecca Nugent and Werner Stuetzle

Abstract

We present a plug-in method for estimating the cluster tree of a density. The method takes advantage of the ability to exactly compute the level sets of a piecewise constant density estimate. We then introduce clustering with confidence, an automatic pruning procedure that assesses significance of splits (and thereby clusteres) in the cluster tree; the only user input is the desired confidence level.