Analysis And Application Of User Phone Behavior Driven By Telecom Big Data

With the expansion of mobile Internet and technological innovation,the number of telecom users is increasing and the scale of business is expanding continuously.The mobile phone number of users has become the identity information of users and has been registered in a large number of Internet applications.Based on the huge telecommunication data set,how to excavate the value behind the large telecommunication data has become an urgent problem to be solved in today's society.Based on the telecommunication data set of telecom operators,this paper analyses and applies the data through the current big data technology and the corresponding strategies.One is based on the historical telecom data to analyze abnormal phone;the other is based on the historical telephone data of users and combined with the browsing information of web pages to mine the target users and carry out accurate marketing.At present,operators at home and abroad have technological innovations in solving the problem of abnormal phone.At present,the abnormal phone technologies such as black-and-white list technology and reputation system technology are all passive detection technologies.Based on the current popular algorithm learning to rank,we analyze and extract the data characteristics,and construct a set of abnormal electricity that can be proactively predicted.In order to improve the accuracy of the model,ensemble learning is used to process the experimental results.Accuracy can reach 86.6% in a certain fraction interval.The three major domestic telecom operators have maintained a high degree of consistency on big data issues.As early as after 2010,they began to focus on developing their own big data business for technological innovation and economic income generation.Based on the resources of big data in telecommunications,through the analysis of user data,the division and classification of users have been carried out,and a lot of development and application have been made in the aspects of telecommunication credit reporting and accurate marketing.Based on the telecom data provided by operators,we try to use multi-objective optimization algorithm to mine the franchise users for a catering franchise company.At the same time,combined with the user's online information,we can increase the quality and accuracy of users and carry out accurate marketing.The accuracy rate reaches 40%.In this paper,two practical problems are analyzed on the basis of large data of telecommunications.For each problem,the model is built by combining the practical machine learning algorithm and intelligent algorithm at home and abroad.The experimental results show that the model is accurate and predictable.At the same time,our model can be packaged and handed to operators for analysis and profit.It has application in telecommunications and commercial fields.