Research And Application Of The Attribute Reduction Algorithm Based On Improved Discernibility Matrix

As information technology and database technology developing rapidly, every day people face a huge amount of data in which there are a lot of redundant data affect our decision-making. Rough set theory is the most effective tool in reasoning and in our extraction of the decision rules.This paper mainly studies the rough set reduction algorithm based on the improved discernibility matrix. The time and space complexity of the traditional method of reduction are high, so this paper improves the traditional method from theoretical and practical aspects of the reduction algorithm.(1)The general attribute reduction algorithm was analyzed. Then this article gives an improved reduction algorithm for the irrationality to deal with the inconsistent decision table.(2)This article analysis the improved discernibility matrix reduction algorithms proposed by the previous,then found that as all the algorithms must scan every record many times, the time and space complexity are high.Based this defect,the author gives the improved algorithms.Through improment,the time and space complexity can reduce.(3)After the attribute reduction, reduction of property values is improved,which can make the decision-making knowledge table most simplified.(4)Through the improvement above,this article gives a general improved algorithm.Through a experiment ,this improved algorithm is proved to be effective.(5)The proposed improved reduction algorithm in this paper will be applied in the field of the mine fan fault.The core idea of this part is reducing the decision knowledge of mine fan fault using the improved reduction algorithm proposed in this paper.