Application And Research Of Data Mining In Telecom Based On Classification Decision Tree And Apriori Algorithm

With the development of telecommunication enterprises and the criterion of telecommunication markets, the information industries department appears to stipulate in February, 2008: there must be no more than 10 types of local price package services. But telecom operators have tens or hundreds of price package services at present, which is inconvenient for consumers to choice and use. As the important product of the telecommunication enterprises, price package service is the bridge between enterprise and customers. So the analysis to it is valued by enterprises. In order to reduce the risk it carries, the price package service will be previewed before carried out. At present the preview often takes "Stimulated Outgo Method", which has following limitations: choosing the customers sample with blindness and unable to know aimed customers of price package services and anticipated month incomes. So customer sample is not related with price package service, and the total index obtained is valueless. The article thinks inversely. First, separate customers according to brand, on this foundation, using association rules algorithm to discover expense rules of the group, then we can work out the corresponding price package services. So once the new price package services are performed, they have meet the rules of market segmentation. Under the premise that the types of price package service are limited, the method can effectively avoid blindness and reduce risks.At present domestic telecom operators have almost completed the design and construction of enterprise database. Based on the database, after the data access, the pretreatment and the transformation, the article has established "the small data warehouse" facing the specific subject, namely the data mart, which alleviates the bottleneck to data warehouse to some degree.The article works on the relative knowledge of data warehouse and data mining, especially the ID3 algorithm of Classification Decision Tree and the Apriori algorithm of association rules, finds out the related information about making price package services, uses the opening data mining tool of Weka to mine data, has discovered the relations between SMS,Tel city,Long-distance telephone and Value-added business, modified the existing price package service.The article provides a scientific basis and breaks the new situation for developing price package service.