— In this paper, we present a novel approach to partitioning pattern spaces using a multiobjective genetic algorithm for identifying (near-)optimal subspaces for hierarchical lea...
Abstract. This paper studies the properties and performance of models for estimating local probability distributions which are used as components of larger probabilistic systems â€...
Kristina Toutanova, Mark Mitchell, Christopher D. ...
Machine learning techniques such as tree induction have become accepted tools for developing generalisations of large data sets, typically for use with production rule systems in p...
Information available in the Internet is frequently supplied simply as plain ascii text, structured according to orthographic and semantic conventions. Traditional document classi...
Classification, which involves finding rules that partition a given da.ta set into disjoint groups, is one class of data mining problems. Approaches proposed so far for mining cla...