Application of the BP neural network classification model in the value-added services of telecom customers

Currently, telecom value-added services are one of the main sources of revenue for China telecom and its operators. While, they lack of effective data mining methods because of they still simply and blindly relying on users' free experience, telephone interviews and other traditional marketing tools. In this paper, we carry out a new analysis in statistical sence simply which applies the fitting method of cubic spline interpolation to describe the relationship between variables and target attribute. According to this analysis, a BP neural network classification model is established for the value-added services of the telecom customers. This classification model is validated for some real samples of value-added services and achieved an average classification accuracy of 89.25%. The result demonstrates that the method of BP neural network classification can well assist operators to make a decision targeting and target customer groups for value-added service.