Attribute reduction method based on information entropy and discernibility matrix
For the information system without decision attributes,the difference matrix without decision attributes is defined,and a"empty invariant"reduction method without decision attributes is proposed by u-sing the calculation method of information entropy under the quotient space theory.Starting from the essence of attribute reduction,the probability of attribute causing the change of the equivalent class is quantified,and then the attribute reduction is realized.The effectiveness of the method is verified by examples,and the com-parison with the existing methods and the analysis of experimental results further confirm that the method can more accurately identify the key attributes and complete the attribute reduction.
without decision attributesinformation entropy under quotient space theorydifferential matrixattribute reduction
万方数据
维普
