Font Size: a A A

Analysis And Research On Traffic Management Of Telecom Operators Based On Data Mining

Posted on:2018-12-29Degree:MasterType:Thesis
Country:ChinaCandidate:R GongFull Text:PDF
GTID:2359330536479772Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of mobile Internet,the revenue of traditional voice and SMS business decrease significantly,and a variety of mobile applications develop so fast,mobile data traffic is also growing rapidly.At the same time,operators are facing the threat of being pipelined by Internet company and the problem that traffic income doesn't increase.In such case,it is particularly important for operators to do a good job in traffic management.This paper analyzes the advantages and disadvantages of the operators' traffic management,and analyzes the user traffic data by establishing the data mining model.Based on the data of Nanjing users' traffic in June 2015,the idea of remote data collection,local environment construction and local analysis of user traffic data is adopted.The paper analyzes the traffic management from three aspects: customer segmentation,Internet preference division,the search of association rules.Customer segmentation mainly selects customer traffic behavior,consumption characteristics as the segment indicators,and takes the k-means algorithm.Combined with business knowledge,users are divided into three clusters: high level cluster,medium level clusters,low level cluster;Internet preference division takes the method of separating the high frequency preferences and general preferences.The two parts are clustered separately.And the result of high preferences part is 4 kinds of single preference clusters and 1 kind of multi-preference cluster,the general part is 9 kinds of single preference clusters,8 kinds of multi-preference clusters;then the terminal information is added to the result data and the FP-growth algorithm is used to mine the strong rules.It can get 16 valuable rules.At last,the paper concludes that the multi-preference users of news,video and finance are inclined to become low level customers;high level users prefer to browse the contents of technology,reading and living service;users of using 4G Huawei mobile phone are inclined to be 4G users.The research of this paper provides the analysis ideas for the operators' traffic management.After improving the system and completing the interface,it can be used as an auxiliary tool for operators to carry out targeted marketing in the traffic management.
Keywords/Search Tags:operators, data traffic management, data mining, WEKA, k-means, FP-growth
PDF Full Text Request
Related items