Power Customer Credit Analysis and Decision Making Using Tolerant Rough Sets Model

Being the deregulation of power market, it is important to evaluate the credit of the participators especial to different kinds of customers. However, there are several factors which are complicated associated with credit of power users. Moreover, the factors are interrelated so that it becomes difficult to design an efficient method to search out independent factors and decision rules supporting power enterprise´ s market strategies. In this paper, a rule mining model using tolerance rough sets is presented in details. Due to the capability of reduction from rough sets based, model in this paper can extract the core rules and factors from customers´ management system. Genetic algorithm is used to determine the optional similarity threshold values among objects and weight of attributes respectively. Experiments have been conducted on real data, and results show that this framework may improve the effectiveness under lower computing complexity.