K-means Clustering-based Fusion Algorithm And The Mobile Customer Segmentation

Data mining technology are widely used in various industries such as medical data analysis, audit data analysis, financial data analysis, and in these areas data mining has made a good application of the results.With the rapid development of the telecommunication market, every company provides an individual service is very important. So how to find the customers' features, how to provide the service to customer, how to use data mining technology to analyze the customer actions has become a hot research area.Clustering analysis is the core technique of data mining, and begin to be used in Telecommunication. After analyzing the traditional clustering algorithms, the paper presents a new clustering ensemble method based on K-means to cluster data. It connects the data points by K-means cluster results, and generates an ensemble function based on threshold, the new algorithm is called K-means-CE. And the paper presents a way to optimal the ensemble clustering results. Some experiments have been done with the dataset in UCI by the K-means-CE. The experimental results show the algorithm is efficient and useful.The paper uses K-means-CE to analyze the telecommunication customer action. According to the results, the customers are category as seven groups, such as high quality customer group, local area customer group and so on. After that, we use a group service as an example, first the relationship between the customer characteristics and group service, and then some analysis have been done to find what kind of service is fit to the group customer.According to the application in Telecommunication Corporation, the application results have been shown the algorithm can be used in the area efficiently. The result of experiment is analyzed in several aspects, and then makes the decision of the market strategy for each category, so that enterprises can better grasp of market dynamics and give effective technical support for mining the potential customers. The experimental result confirms the validity of the clustering algorithm.