Affiliations: [a] Department of Computer Science and Engineering, Sri Ramakrishna Engineering College, Coimbatore, Tamilnadu, India
| [b]
Nehru Institute of Technology, Coimbatore, Tamilnadu, India
Correspondence:
[*]
Corresponding author. R. Anuradha, Assistant Professor, Department of Computer Science and Engineering, Sri Ramakrishna Engineering College, Coimbatore 640122, Tamilnadu, India. Tel.: +91 422 2460088; Fax: +91 422 2461089; E-mail: [email protected].
Abstract: Machine-learning and data-mining techniques have been developed to turn data into useful task-oriented knowledge. The algorithms for mining association rules identify relationships among transactions using binary values and find rules at a single-concept level or multiple levels. Mining associations among itemsets only by using support and confidence thresholds at different levels of hierarchical data would not give interesting rules both for binary or quantitative data. This paper proposes a two phase algorithm that mines rare generalized fuzzy coherent rules at inter-cross level hierarchies. During phase-I both positive and negative fuzzy coherent rules are mined and in Phase-II, rare generalized fuzzy coherent rules are extracted from the resultant rules obtained from Phase-I. The algorithm framework works on top down methodology in generating positive and negative fuzzy coherent rules and mining rare generalized rules from it. Experiments conducted using synthetic dataset show the performance of the proposed algorithm in terms of the number of rare generalized rules generated, compared to fuzzy multiple-level association rule mining algorithm.
