Exploring Similarity Relations According to Different Contexts in Mining Generalized Association Rules

Taxonomies are used in different steps of the mining process, but in the generalization they are mainly used during the pre-processing. In the same way, in fuzzy contexts fuzzy taxonomies are also used, mainly, in the pre-processing step, which is the generation of extended transactions. Besides, it is possible to see that many works have explored mining fuzzy rules and linguistic terms, but few works have explored different steps of the mining process. Moreover, questions like semantic enrichment of the rules have been little explored. In this sense, this work presents Context FOntGAR algorithm, which is an algorithm for mining generalized association rules under all levels of fuzzy concept ontologies. Besides, the exploring of rules containing similarity relations according to different contexts will be introduced. In this work the generalization is done during the post-processing step. Other relevant points are the specification of a generalization approach; including a grouping rules treatment, and an efficient way of calculating both support and confidence of generalized rules.