The Research Of Data Mining In Mobile Communication Enterprise Based On Decision Tree

The technology of Data mining means to find connotative, novel and useful information out of database, and it is widely used in many realms. Classify and prediction is the main measures in Data Mining, which make great progress in theory and method till now, and Decision Tree arithmetic is the symbol. Based on the increasing demand of the instruction of Data Compression Enterprises database and the demand to business intelligence, the application of data mining is urgent, and the Decision Tree are also facing a challengeable future.The thesis focuses on the characteristics of the mobile communication enter- prises and the demand to maintain the customers, to study and provide a way to refine the Decision Tree. It is applied to the analysis to customer churn and incoming telegram reminding, etc. The works done are followed:1. The essay describes the actuality and development of Data Mining technology, and the concept, main content, application of Data Mining in mobile telecom realm en are introduced.2. The thesis describes the Decision Tree of Data Mining technology and introduces the classic Data Mining algorithm-ID3 algorithmic in Decision Tree. And also analyzing and summarizing the quality, function and characteristic of ID3 algorithmic. Aiming at the shortcoming of traditional ID3 algorithmic which takes a lot of time calculating and attributes tendency, we come up with an improved ID3 algorithmic. It can converse the entropy function to table calculation based on data by successive feature. So we improve the efficiency. The algorithmic make use of attribute weighted factor to get over the attributes tendency.3. Using improved Decision Tree algorithm to establish the customer churn prediction model of a mobile telecom enterprise. And applying it to decision support such as: churn prediction, the reason of customer churn analyzing, customer detainment and so on.4. Using improved ID3 Decision Tree algorithmic to establish the target customer prediction model in the incoming telegram reminding of a mobile telecom enterprise, moreover, we commercially describe and analyze the prediction result, and provide effective decision support for business marketing strategy in a mobile telecom enterprise.