Owledge Reduction in Database Systems

In the rough set model based on database systems, the algorithms of calculating cores and reducts are very efficient and scalable in data mining applications. However, one serious drawback of these algorithms is that they are only applicable for consistent decision tables. In this paper we propose an algorithm which is able to receive consistent decision tables by initial decision tables while at same time preserve both cores and reducts, thus overcome this drawback. By utilizing set-oriented operations, this algorithm is rewritten in database systems. Analysis shows that calculating cores and reducts of inconsistent decision tables is still efficient and scalable when this rewritten version is employed