| Today,4 g network is all over the country, the operators are vigorously promoting its 4 g network, and the 4 g terminals are becoming operators’ new bestow favor on, they want to take this opportunity to tout its 4 g network, expend the 4 g market, promote the users’ value and add more revenue. This paper, by analyzing the information such as Shenzhen telecom users’ ARPU, voice, the internet traffic and terminals and using logistic regression models identify potential changing phones of users, then using association rules to haven changed phone of users’ former age, terminal brand, price, calling behavior and the later terminal brands and models, it is concluded that replacement rule, predict which terminals, users may change.At the same time according to user’s existing packages category, we can choose to suit the user’s 4 g package, eventually targeted marketing, let the customers experience "one-stop" work style service. Shenzhen marketing model according to user’s potential behaviors will draw 150000 list, and matching the user could replace the brand and model, outbound personnel for outbound, the shooting can achieve 26.3%, greatly improving the original transfer rate is about 5%, it makes Shenzhen telecom company expanding 4 g users in a short time, promotes the user value. This paper, by using data analysis and mining knowledge applied in the actual business, gain higher practical value, show the charm of statistics and meaning!... |