| The paper take XX telecommunication business system as background, take service churn as object. The paper take XX telecommunication business analysis system as a background, take the telecommunication service churn as the object of study, carries on the analysis to the service churn. The service churn is not the true .It was only the user in the net has used other operation business service, created the service churn in this net. Analyzes the behavior of consumer before service churn, found their rule of churn, adopt the essential measure, avoid their service churn. The paper based on this foundation, carries on analyzes to the service churn.The paper first construct the data warehouse, the aim for the construction the data warehouse is OLAP analyzes and providing the data support. You can learn the general understanding through the OLAP analysis to the service churn, and has provided the data support for the service churn further analysis. The following is the data preparation, it including the service churn coefficient computation and clustering analysis. It is the powerful data support for the service churn model establishment. Establishment the service churn model is key point of the paper. According to the topic characteristic of paper research established the LOGSTIC regression model of service churn. And compute the model coefficient with the SAS, latter carries on the WALD method and the ROC curve model exam this model. Finally, uses the test data verify the model, and gain the model grading.The paper adopts several methods of the multitudinous statistical analysis reorganization data and the building of the service churn. Simultaneously also applied to mathematics statistics method verify the model, and using the theory unified the actual appraise to the model, had proven the paper established the model of service churn is significance. It will have been the practical value and the reference value to the servicechurn for telecommunication. |
PDF Full Text Request