Host: Japan Society for Fuzzy Theory and Intelligent Informatics (SOFT)
We propose a heuristic attribute reduction algorithm based on random selections of attributes. Rough set theory proposed by Pawlak provides theoretical foundations of set-theoretical approximation of concepts and logical reasoning from data, and attribute reduction is one of the most important research topics in the aspect of reasoning from data. The computational complexity of computing all reducts from the given data is NP-hard and there have been many heuristic attribute reduction algorithms; however, almost proposals compute just one candidate of reducts from the given data. In this paper, we propose a heuristic attribute reduction algorithm to compute reducts as many as possible, which is based on construction of small-sized decision tables from the given data by using random selection of attributes.
