Application of variable precision rough set attribute reduction algorithm

In order to remove the drawbacks of the classical attribute reduction algorithm based on attribute importance which can´t ensure to gain the simplest decision table and the best noise resisting ability, this paper presents a improved attribute reduction algorithm based on β - variable precision rough sets. A parameter β represents the error resolution and makes the decision table more simple and reliable. Compared with the classical attribution reduction algorithm, β - variable precision rough sets attribute reduction algorithm has better generalization and ability of resisting noise. These two algorithms are used on the simulation of Car Test Results and the results verify the superiority of the improved algorithm.