Non-leap subspace clustering on categorical data
Min Wang · Jisuanji gongcheng yu sheji · 2009
Clustering,in data mining,is useful to discover distribution patterns in the underlying data.With the growing demand on cluster analysis,a handful of clustering algorithms are developed.Existing subspace clustering algorithms for handing high-dimensional data focus on numerical dimensions,but not on categorical dimensions.Since categorical data have not the inherent distance function as the similarity measure,traditional cluster validation techniques based on the geometry shape and density distribution cannot be applied to answer this question.The entropy property of the categorical data is investigated for determining a set of candidate centroid and a non-leap subspace clustering method is proposed to get the subspace associated with each cluster,then some cluster memberships changing rules using the objective function is deduced.Finally,some experiments are provided to show the effectiveness of the proposed method.