Towards a semantic and statistical selection of association rules

摘要: The increasing growth of databases raises an urgent need for more accurate methods to better understand the stored data. In this scope, association rules were extensively used analysis and comprehension huge amounts However, number generated is too large be efficiently analyzed explored in any further process. Association selection a classical topic address issue, yet, new innovated approaches are required order provide help decision makers. Hence, many interesting- ness measures have been defined statistically evaluate filter rules. these present two major problems. On one hand, they do not allow eliminating irrelevant rules, on other their abun- dance leads heterogeneity evaluation results which confusion making. paper, we propose two-winged approach select in- teresting semantically incomparable Our statis- tical helps discovering interesting without favoring or excluding measure. semantic comparability decide if considered related i.e comparable. outcomes our experiments real datasets show promising terms reduction