SEMMDPREF: algorithm to filter and sort rules using a semantically based ontology technique

Decision Support System designs the means and methods to go through a series of processes with the purpose of satisfying the needs of the decision-makers. Most of the existing algorithms mining rules usually produce a large number of rules suffering from the problems of thresholding, redundancy, and overlapping. There is likelihood that some of these rules are already known and hence trivial and some may be meaningless altogether. To tackle these problems, this paper suggests an approach to discover interesting rules by pruning and filtering them. The approach consists of introducing methods and techniques based on semantic significance, the notion of dominance between rules and userpreference. Our approach neither favors nor excludes any measures. More importantly, specifications of threshold are easier to deal with. Concerning algorithm evaluation, we use a real database, and we compare our results with others of other algorithms such as the Most Dominant and Preferential Rules:MDPREFR.