A Customer Segmentation Research Based On Data Mining In Telecommunications Business

Owing to intense competence in international telecommunication market and the development of the market itself, traditional domestic business model has gradually become the one based on customers and data. Therefore, customer segmentation is the prerequisite and statistics based system, which does not facilitate the information interchange between customers and enterprises. As a result, these methods are far behind meeting the requirement of complex business analysis. Along with the depth spread of data mining among telecommunication enterprise management, it is becoming urgent and critical to implement multi-level, multi-latitudes and focused customer segmentation by adopting various advanced data mining methods.In this thesis, the data mining technology in telecommunication industry has been researched. Through investigations of the telecommunication industry customer segmentation model and its algorithms, we aim to propose a multi-dimensional customer segmentation method, which can be used to guide telecom companies to take different measures for different customers. The main contents and fruits of the thesis are outlined as follows:1. Researched on the basic process and the data pre-processing technologies of data mining, as well as several current more sophisticated data mining technology, determined the data mining method used in telecommunication industry customer segmentation.2. Described the Guangdong telecommunication data mining environment and the application, and designed the solution of customer segmentation based on the Guangdong telecommunication's facility environment.3. Gave concrete steps of customer segmentation data pre-processing, including setting up the customer analytic record (CAR) model and the enterprise logical data model, data integration and clean, and the CAR design.4. Realize a customer segmentation model based on k-means clustering by Clementine.