Measuring customer service satisfactions using fuzzy artificial neural network with two-phase genetic algorithm

Abstract

In this chapter, we propose a new method based on genetic algorithms (GAs) for fuzzy artificial neural network (FANN) learning to improve its accuracy in measuring customer service satisfaction for establishing a principle of ecnomical survival in business area. The analysis is based on linguistic values received from customer service satisfactions index where fuzzy modeling, as one of possible ways, has been used to process these values. Here, customer's satisfaction is considered as a key factor for the analysis based on his/her preference as the scope of qualification for organization service. In the proposed method, we have introduced two-phase GAs-based learning for FANNs. In the neural network, inputs and weights are assumed to be fuzzy numbers on the set of all real numbers. The optimization ability of GA is used to tune alpha-cuts boundaries of membership functions for fuzzy weights. Here, five alpha-cuts are used for tuning as other researchers have used, which in two-phase method; two of them are for first phase and three of them for second phase. This leads to obtain better results for FANN. Comparisons are included with another method using two data sets to give some analyses to show the superiority of proposed method in term of generated error and executed time.From the experiments, the proposed approach has been able to predict quality values of possible strategies according to customer's preference. Finally, the ability of this system in recognizing customer's preference has been tested using some new assumed services.