The Application Of Data Mining Technology In The Stock Of Telecom Broadband Business

With the decline in the development of broadband technology and broadbandtariffs, and vigorously promote the telecom package, the broadband market shift fromthe incremental development model for the stock of development model, in the caseof stock patterns, reduce churn, extend broadband customers product life cycle, toprevent his net instigation to become the stock of the development of telecombroadband a top priority. In this case, telecommunications companies must establish acustomer-centric "management model. Using data mining technology on massivetelecom enterprise customer data mining and analysis, from the loss of customers toexplore the loss characteristics of these features to existing customers, existingcustomers, the loss of early warning, is to enhance the operation of the stock oftelecom broadband an important method of great power, significance and applicationvalue. This paper uses a combination of theoretical analysis and empirical research,conducted in-depth study of data mining in a number of specific applications intelecommunications in the stock business. The main content is as follows:1. Detailed analysis of the status of the stock of telecommunications enterprisebroadband business Telecom broadband customer-oriented theme to the loss of earlywarning data warehouse architecture, loss warning theme for the telecom businessbroadband for a detailed analysis, and design based on the loss of a scene the datamodel, and model the effects of assessment and evaluation.2. Systematic introduction to the theory of customer churn, and the theory ofpractice. Paper loss warning to a city telecommunications broadband project as thecarrier, the early warning system designed to the customer as the center, under andloss of retreat feedback Telecom Broadband. The main line to Telecom Broadbandcustomer life cycle management theory, the loss of customers to the theory into thecustomer life cycle, to maintain the retention of the customer life cycle characteristics,and given for the maintenance of marketing tactics and scripts. 3.Improve classification model based on C5.0decision tree algorithm, asupervised classification model evaluation methods, and predict the results of severalmodel training, the final output to meet business goals and quality constructiontemplate model, so that broadband loss warning system has a high value.This research work as part of the telecommunications industry, businessintelligence systems, loss of customers through the application of theory and datamining technology, with a useful reference for research and development and practicalapplication of the loss of telecom broadband system, has an important significance intheoretical research and engineering practice...