Application research of improved neural network in customer value classification for logistics enterprises

Correctively classifying customer is a key issue for logistics enterprises to realize customer relationship management (CRM) effectively and is becoming a hot research field for the unit and researcher related. A novel customer value classification for logistics enterprises is advanced based on BP neural network algorithm. First, a customer value evaluation indicators system for logistics enterprises including customers' present value indicators and customers' potential value indicators is designed. Second considering that BP neural network algorithm has high classification accuracy but low convergence, fourier basis functions is used to improve traditional BPNN algorithm to speed up model convergence and to simplify model structure. Finally the simulation results shows that not only the problem of convergence speed has been solved, but also the simplicity of the model structure and the accuracy of the classification are ensured when the new algorithm are used in electronic customer relationship management practically.