The Improved BIRCH Algorithm’s Application In The Customer Segmentation Of The Communications Industry

With the growing competition in telecom market, by classifying customers to provide different services to different customers, has become the focus of attention of telecommunications enterprises. With the extensive application of information technology, telecommunications enterprises have accumulated large amounts of customer data. Faced with massive data, traditional customer segmentation methods have been unable to meet the needs of enterprises. Using data mining technology for deeper customer segmentation is very necessary and urgent.The BIRCH algorithm of clustering technology and application of BIRCH algorithm in customer segmentation of communication industry are studied in this thesis. Major works are as follows:Firstly, the BIRCH algorithm of the clustering analysis is studied in this thesis. The BIRCH algorithm ideas and basic processes are described. Advantages and disadvantages of the algorithm are analyzed. Aiming original BIRCH algorithm’s problems, the corresponding improvement methods are given:using log likehood distance to replace euclidean distance in original BIRCH algorithm, using agglomerative hierarchical clustering combing bayesian information criterions to carry out the second clustering and determine the number of clustering. Then, the simulation experiments demonstrate that the improved algorithm is good at clustering stability and the time efficiency.Finally, the improved BIRCH algorithm is applied to the telecom customer segmentation. In this thesis, we get ideal result and analyze this result. So marketing strategies are set up to different customer groups and reasonable decision supports are provided for the enterprise. Thus, the subject has a certain practical significance and application value.