Integration Of Marketing Research And Data Mining On Churn Management

Since the data mining technology invention, go by a period of rapid development, the technology has received worldwide attention from many industry, and has formed many successful case of data mining technology in the production, to improve the business development, greatly improve the competitiveness of enterprises. Churn management is such a typical case of data mining in the telecom industry, many experts and scholars carry out research, have accumulated considerable experience.In this paper, a broadband churn management task as the breakthrough point, the data mining technology in the process of practice and application problems are analyzed and discussed. This paper will introduce another common method of dealing with the loss of customers in the development process, namely the marketing research, this method is a belonging to the marketing discipline, on the business problem analysis and solving a set of complete method system, after analysis, marketing research and data mining is one of can be complementary method. This paper will put forward into the data mining solution with the concept of marketing research, so that the two can complement each other.In addition, classification of imbalanced data set is a problem of churn model, the paper will introduce a parameter named misclassification cost, and provides a method for obtaining configuration through the evaluation matrix of optimum parameters, to avoid overtraining.Then this paper through an actual case, churn analysis of broadband users, verify the solution in practice.At the end of this paper will summarize the key points to the successful implementation of the case, and the prospect of the solution in other business problems or industry.