A knowledge granularity based heuristic algorithm for attribute reduction

Rough set theory is a new mathematical tool to deal with the imprecise, incomplete and inconsistent data. Attribute reduction is one of important parts in rough set theory. Currently, lots of literatures have proposed many algorithms for attribute reduction based on similarity. But all these algorithms just consider the connection of condition attributes and decision attributes, and the similarity of condition attributes is neglected. A heuristic algorithm for attribute reduction based on knowledge granularity is proposed. Firstly, we calculate the similarity between condition attribute and decision attribute, and then use the similarity between different conditions attributes to measure and choose important attributes which are added to the reduction set. Theoretical analysis and experiments show that the algorithm of this paper is efficient and feasible.