Research And Analysis Of 4G Charges Based On Big Data

As of July 2016,China's population reached 1.383 billion,and the total number of mobile phone users reached 1.321 billion,with the number of 4G users reaching 714 million.2G,3G and 4G mobile users' total number is close to saturation.The incremental growth model that simply increases the number of 4G mobile subscribers has failed to meet the needs of mobile operators.Operators must focus on the stock growth model and make it become a most important development model.This is of great significance to the business growth of the operators and the economic construction of our communication sector.This article analyzes the user data of a province telecom operator by using the decision tree,random forests,AdaBoost,support vector machine(SVM)and K-means algorithm,from "potential 4G users model" and "4G user charges characteristic model" two aspects to analyze the characteristics of the 4G users and charges.First of all,through the exploratory analysis of the basic characteristics of the 4G users(gender,age,etc.)and consumption characteristics(ARPU,DOU,MOU,etc)macro understanding.Then,analyze user data by using the data mining technology.In the model of mining potential 4G users,support vector machine(SVM)algorithm of accuracy has a huge advantage.The accuracy rate is improved by using logarithmic transformation of the variable.In 4G user charges characteristic model,based on user DOU,MOU to cluster analysis,two variables for the user to grasp the overall consumption characteristics,it is concluded that "balanced","data flow driven","the voice driven" three types of users.Through the comparison of different gender and overall clustering characteristics,it is concluded that k = 13 on the degree of classification user rates and the best cost is low.Make Suggestions according to the result of clustering analysis for the Telecom operators.