Ontology-Based Association Rule Quality Evaluation Using Information Theory
Xia Shi Xiong, Fan Li, Zhang Lei · 2010
Support and confidence are two main parameters of association rule mining, the first is used to measure the statistics importance of association rule, and the second is used to measure the reliability of association rule. The quality of association rule does not have quantitative evaluation criterion. In this paper, Quality index is proposed, the subjective and objective aspects are integrated and information theory is introduced in order to evaluate multi-level association rule's quality based on domain ontology. The quality index of rule can be an important reference in redundancy treatment and rule application. Finally, the experiment shows one of the applications of quality index in multi-level association rule mining and redundancy treatment ontology-based.