A new algorithm for electronic customer relationship management

This paper presents a new customer classification algorithm for electronic customer relationship management. First, based on consumer characteristics and behaviors analysis, 21 customer classification indicators are designed, including customer characteristics type variables and customer behaviors type variables. Then, when considering the advantages of high classification accuracy of BP neural network model, and in order to speed up the convergence of the model, the paper constructs a new Legendre wavelets neural network model to classify customer for electronic customer relationship management. Finally the experimental results verify not only the problem of convergence speed has been solved, but also the simplicity of the model structure and the accuracy of the transformation are ensured when the new algorithm are used in electronic customer relationship management practically.