Hybrid Filter for Attributes Reduction in Soft Set

Abstract

The purpose of this research is to overcome parameterization reduction limitation that focuses only on individual parameter reduction, whereas in some cases the individual parameter reduction is not sufficient even implies reduction. It was found that the dimensions sometimes are not able to reduce the number of data in the case of big data; hence, for this reason it became necessary to look for an alternative technique that can significantly reduce the parameters. This paper proposed an alternative decision partition order method based on rough set indiscernibility to select attributes reductions in soft set using decompositions. For significant candidates, the method decomposition partition order used R supp checking to confirm the correctness of the reduction. Comparison of the reduction methods shows that the proposed method provides better result than the parameterization reduction in enhancing reduction. The false candidates were filtered in the huge candidate reduction by the Min supp. The proposed method can be used to maintain object before attribute reduction as well as to reduce parameter size drastically while maintaining consistency in decision making.