Abstract: Data-mining techniques are developed to transform raw data into suitable knowledge-oriented data. The algorithms for mining association rules identify relationships among transactions using interesting measures like support and confidence at a single-concept level or multiple levels. Using support and confidence alone for mining associations would not give interesting rules both for quantitative as well as binary data. This study proposes a fuzzy coherent rule mining algorithm at multi-level hierarchies to discover the significant rules in quantitative transactions. The proposed method combines fuzzy coherent rules mining concept with that of taxonomical mining in a quantitative database. The algorithm works on a top down methodology in traversing the data that exists in a hierarchical form. An experimental comparison with the fuzzy coherent rule mining methodology conveys the significance of the proposed algorithm in finding the level-wise coherent rules.
R. Anuradha and N. Rajkumar, 2016. Mining a Complete Set of Fuzzy Multiple-Level Coherent Rules. Asian Journal of Information Technology, 15: 3441-3448.