Forecast of importance weights of customer requirements based on artificial immune system and least square support vector machine

With view to satisfying customers, it is important to correctly ratify importance weights of customer requirements in quality function deployment (QFD). The twenty-first century is marked by fast evolution of customer tastes and needs. Customer requirements could vary with time, customers´ preferences and competitive ability of product manufactures. It is urgent and critical to capture the dynamic customer requirements for new product design in QFD. To provide an effective method to predict the importance weights of customer requirements, the model was proposed for forecast of importance weights of customer requirements based on least square support vector machine (LSSVM). To acquire the better parameters of LSSVM, artificial immune system was used to optimize the parameters of LSSVM and the AIS based LSSVM was proposed for forecast of importance weights of customer requirements. To verify the approach, a case was used by comparison between AIS-LSSVM and LSSVM in this paper. The result showed the LSSVM optimized by AIS had better performance than the LSSVM without parameters of optimization by AIS.