Design And Realization Of The Segmentation Module Of Communication Company

Owning huge quantity of users is one of the important characteristics of Chinese telecom transaction companies. After going through the initial phase that the transaction and business were rapidly developed, the quantity-increase market of telecom market is shriveled while the quantity-existing market is saturated day by day; meanwhile, transition from the previous extensive operation model to the currently precise sale model is urgently needed , which will transfer the business operation type from concentrating on the product and service to the one concentrating on the needs of customers, among it, the key point is : be able to conduct analysis and research on the huge-quantity users group by means of close classification, understand behaviors characters and varieties of users' consumption, thus makes the differentiated sale operations available. Being aimed at data- mining technology, this article implements a certain analysis and research on the application of how to conduct module-establishment and analysis by means of clustering for the telecom huge-quantity users.Some of basic knowledge concerning data -base is firstly presented in thisarticle as well as the contents and classifications of the previous telecom transaction companies' business data are simultaneously introduced. Based on the engineering practices, especially focused on the data-module hierarchy of data base, a certain description and discussion & study are performed.Furthermore, some of the basic conceptions and technologies about data-mining are also introduced in the article; meanwhile, the merging analysis module adopted for this time's data clustering is also described by details, especially for the K-means as well as the features that are directly used by this engineering; simultaneously, the CRISP-DM flow process is also presented for discussion due to the reason which has been introduced for this time's module-establish by means of close-classification.According to the above-mentioned standard flow process, this article provides detailed introduction of the whole process adopted by one telecom transaction company to establish whose close classification for the users based on the actual business data. For the mainly involved six phases such as business understanding, data understanding, data prepare, module establishment, evaluation and data release, one-by-one actual study & discussion are conducted while combining with various actual application examples of data analysis, by which, actual application of the data mining in the business service is comparatively and actually revealed. Meanwhile, analysis and study on the partial details among which are conducted in a certaindegree, for the user groups obtained from the clustering, research and analysis are also done while linked with the telecom business.At last, the article furnishes presentation about: after the clustering-module is established and the data mining result is available, how will the telecom transaction companies apply the knowledge discovered from the data mining into the actual business operation as to improve the efficiency and quality of the typical sale, based on understanding the characteristics of these clustering user groups aswell as acknowledging the ideas and suggestions for the differentiated sale services.