Prediction Of Mobile User Terminal Changes

In the future of the telecommunications market,the frequency of the customer changing the phone and the degree of customer relying on the phone will be more and more high.Mastering the trend of customer terminal replacement and accurately predicting the changes in mobile terminal users to effectively develop the policy to prevent the loss of customers has a very important significance for the telecommunication.Because of the advantages of data mining technology in dealing with massive data and the maturity of its algorithms,making data mining technology in mobile terminal users to predict the problem has played a very important role.This paper discusses the application of data mining algorithm in telecom operators from the aspects of telecom operating environment and user behavior attributes of operating enterprises based on the theory and method of data mining and customer value.This article uses the A telecom operator to carry on the data of a certain province customer,and uses the R language data mining tool to carry on the data integration,the data cleansing,the data transformation and so on to the data set.Then,the decision tree and the neural network algorithm are used to train the sample data to establish the prediction model,the results of the two models were evaluated and compared,Accuracy,Recall,and ROC curve.indicates that the decision tree algorithm model is more suitable for prediction studies.Finally,the decision tree model is used to predict the user behavior attributes for the next month and identify the users who are potentially replacing the terminal.As the method to join the "tracking" attributes,such as replacement brand,terminal type,network type,the data set can better predict the mobile terminal user changes,making the predicted hit rate higher.This has overcome the shortcomings of previous research on potential customers of telecommunications,such as the use of customer consumption information,personal information,other information and other raw attribute data,these raw attribute data is difficult to truly reflect the loss of customer behavior.The result show that the model is basically in line with the needs of telecom operators,to provide valuable predictive information to the relevant decision-makers and marketing staff,the telecom operators to develop relevant guidelines have a certain guiding significance.With the continuous development of data mining technology,coupled with the evaluation system,business attributes,forecasting model of continuous improvement,the proposed method proposed in the telecommunications industry will have an increasingly important role.