| With the structure reforms of the supplying side promoted,the accurate description and evaluation of customer value has become the primary task of the telecom operators.However,the traditional methods rely too much on the prior knowledge of experts to deal with the vast amounts of data and unable to handle the imperfection,uncertainty and high dimension of the evaluation system.Therefore,an improved rough set model is introduced to put forward a new method to evaluate telecom customer value.First of all,the characteristics of the industry and influence factors of customer value are analyzed,and a three dimensional evaluation system based on the “current-potential-external” value of telecom customer is put forward to realize the all-round,multi-dimensional evaluation of customer value.In accordance with the principles in the construction of the index system,the representative attributes are chosen as the index in each dimension and the specific way of calculation is given.Secondly,in view of the characteristics of the evaluation system and data,the traditional rough set model is optimized.The limited dominance relation is introduced into weighted multi-granulation rough set.Besides,a new algorithm of attribute reduction based on the constant lower approximation is discussed,and the method of acquiring decision rules is given.Through the analysis of the cultural security warning in the network,the advantages are proved.Then,the telecom customer value evaluation process based on the improved rough set model was proposed.Finally,a group of real customer data of a telecom operator is introduced to complete the empirical analysis,with the key customer attributes and decision rules preliminarily acquired.The result shows that the model is not only applicable to the magnanimity,incompletion and high dimension of the data,but also realizes the customized learning based on the data itself and saves repetition work of the experts.Therefore,in the evaluation of the telecom customer value,it turns out that the model could be regarded as the theory guidance and decision basis. |