Using Extended Multidimensional Pattern Relation for Multidimensional On-line Mining

Abstract

Although incremental data mining and online mining approaches are rather efficient, they usually can not flexibly obtain association rules or patterns from portions of data, diversely consider problems at different aspects, and provide on-line decision supports for users. In the past, we thus developed the multidimensional on-line mining approach to select and integrate related mining information for diverse mining requests. It uses the multidimensional pattern relation to structurally and systematically store the additional context information and mining information for each inserted block of data. It may, however, get loose upper-bound supports of candidate itemsets and thus cause datasets to be frequently re-processed for new mining requests. In this paper, we attempt to apply the concept of negative border to enlarge the mining information in the multidimensional pattern relation to help get tight upper-bound supports of candidate itemsets, and thus reduce the number of candidate itemsets to be considered. Based on the extended multidimensional pattern relation, a corresponding online mining approach is proposed.