Research On Marketing Strategy Of Telecom Users Stability Based On Data Mining

In recent years,with the gradual popularization of 4G technology and the growing influence of national policy "speed down fee",the competition between telecom operators intensified which emerges "customer battle" and "price war" frequently.While the scale of telecom users is gradually saturated,operators are aware of the key point to promote income has changed.We should take more importance of the value of users rather than the scale of users so that the most crucial work for us is to maintain the stability of users.At the same time,some of the marketing strategies,such as low price and high speed of Internet,cause the increasing scale of telecom users.Although the operators apply strategies to keep users,for example sending short messages monthly and giving traffic and data,the cost of investment and performance is not proportional.Therefore,it requires us to predict which kind of users have the risk of off-grid and then analyze the behavior of such users,so that we can take an effective marke ting strategy.In order to complete the establishment of the stability model of the telecom users,this paper has completed the following research on the actual case of a mobile operator in the small city.First of all,comparing data mining and large data technology,we selected the data mining technology for the research according to the size of the data.By reading a lot of related literature on data mining in customer churn management,this paper described three kinds of classification algorithms which are the most widely used,higher precision and better adaptability.These algorithms used in the paper are logistic regression,decision tree and random forest algorithm.We used them to build the forecasting model in this paper respectively.Secondly,this paper using the data of all the users of a small city,selected the variables and established the stability model.By comparing the prediction accuracy,the random forest model was selected as the final model for predicting user stability.According to the model results,we analyzed the variables that affect the stability of the users,and found out some characteristics of off-grid users.Meanwhile we discussed the influence of the existing marketing strategy on the stability of the users.Finally,this paper used the K-means clustering algorithm to classify the customers.By using the forecast results,we marked out three kinds of users who presented differently in stability.In according to analyze the characteristics of the classification users,we made marketing strategies separately based on 4P strategy.