An information mining method for deriving weights from an interval comparison matrix

This paper provides a simple and practical approach for obtaining the weight information from an interval comparison matrix. First, a series of weight vectors are acquired by three linear programming models. Next, the crisp comparison matrix is gained by the convex combination of the vertices. The interval comparison matrix is approximately replaced by the crisp comparison matrix. Then, a method of the deviation degree is used to get the crisp weight vector. Finally, a numerical example and the comparison shows the validity and practicality of our developed method. More potential weight information is mined.