Matrix Algorithm for Quick Attribute Reduction Based on Decision Table

At present,the efficiency of matrix algorithm for attribute reduction is not ideal.Aiming at this situation,in this paper,firstly the concept of the simplified decision table is introduced;a large number of repeated objects in decision table are deleted effectively.At the same time,the construction of the simplified matrix based on the model of positive region is provided,which can reduce the searching space of attribute reduction.Furthermore,it is analyzed theoretically that the attribute reduction definition based on simplified matrix is equal to the attribute reduction definition based on positive region.And an efficient measurement method for attribute importance is presented.On this condition,a matrix algorithm for quick attribute reduction is designed.Finally,an example and experimental comparisons are employed to illustrate the efficiency and feasibility of the algorithm.