Analysis Of Telecommunication Data On The Basis Of Data Mining

With the Chinese telecommunication reform reorganization and the issuing of the3G licenses, the Chinese telecommunications market environment has changedfundamentally. The competition in communications industries and for customers hasbecome increasingly fierce. In the competitive environment of customer-centric,identifying the customer’s consumption characteristics, consumption patterns, andestablishing the accurate marketing model of telecommunications customers’ businessfrom the complicated historical data have become one of the enterprises winning chips.In this paper, data mining techniques was used to analyze the telecommunications data.Firstly, the research background, research status and the basic theoreticalknowledge of data mining were introduced briefly in the paper. And then, the telecomdata was preprocessed according to the characteristics of the telecommunication dataand telecom data set used for data mining analysis in this paper was established. Andcombined with telecom business requirements, the data mining technology was used intelecommunication fields of data analysis.In this paper, the analysis model of telecommunication business downturn wasestablished by cluster analysis and C5.0decision tree method. The model was used topredict customers with the business downturn tendency, and then the appropriateretention measures were made to reduce the loss of telecom operators. For GPRSservices used by customers, the associated analysis model of GPRS services wasestablished using cluster analysis and Apriori association rules, then GPRS associatedservices were recommended to the high-value customers according to the results of theanalysis. For the unbound GPRS services with good development prospect, customerprediction model of telecom unbound GPRS service was established using the CARTdecision tree and C5.0decision tree algorithm respectively, and compared the results of the model, the result of C5.0decision tree model was better than the CART decisiontree model. And then according to the CART decision tree model, businessrecommendation was made to target clients, so as to improve the profit of thetelecommunications operator.