MeAR-CP: Evaluation of the Quality of Association Rules Using Constraint Programming

Association rules mining is one of the most relevant techniques in data mining.Aimed at identifying interesting connections and associations among groups of items or products within extensive transactional databases.However, this technique can yield too many rules, among which some are irrelevant and/or redundant.Thus, may present obstacles for the decision-maker.This Highlights the importance and challenge of evaluating extracted knowledge to define the most interesting association rules.In order to address this issue, we presented a constraint programming approach to evaluate the relevance of association rules.Our MeAR-CP approach involves filtering association rules using metrics such as IR, Cosine, Lift, Kulc as constraints solved by Choco constrain programming tools, and proposed our metric called 'Score'.The experiments are conducted on various datasets from the UCI Machine Learning Repository.We evaluate both time and rules.The results obtained from our experiments underscore the effectiveness of our approach in reducing irrelevant and redundant rules within an effective timeframe.