Customer Churn And Network Information Dissemination Model Based On Data Mining

At present, the communication industry development speed is make a spurt of progress,and the telecommunications industry is gradually opening, competition between domestic and international telecommunications is increasingly fierce, such as the fight for customers, expand market share and so on, for now, marketing has become more and more mature. The instability of customer groups drove them more biased in favor of people sense to market their interests, change and adjustment of the market are closely related to the customer, how to enhance the customer loyalty, has become a research hotspot for modern marketing personnel, Research on nine USA industries indicate that old customer retention rate increases 5%, the enterprise will increase the profit of 25%~85%.At the same time, it will spend 5-7 times of the cost to produce a new user than to keep an old customer, however, the success rate is 16 times higher to retain old customers than to generate new customers. Therefore, to reduce the loss of customers also means reducing the cost and loss of profits, early detection of the loss potential customers, and to take effective measures to retain customers is the urgent need to solve the problem of telecom industry.However, in the limited time, how to analysis of the massive operation data effectively to make the results have guiding significance, and discover the loss of customers, to take effective measures to retain and maintain existing customers is worthy of further study and discussion. Data mining is also considered KDD, KDD(Knowledge Discovery in Databases-KDD: knowledge discovery) that is knowledge discovery based on database, refers to the extraction of interesting knowledge from large databases or data warehouses, such knowledge is implicit, previously unknown, potentially useful, easy to understand information. Economic globalization makes increasingly fierce market competition to all walks of life in the, for the enterprises in information age, how to find and make full use of the massive data of implicit knowledge has become the key factor to seize the opportunity to improve the core competitiveness. Customers significantly determines the success of the enterprise and profit, the application of data mining technology in the customer management provides the basis for the enterprises management and distribution to make full use of existing resources to maximize profit.In this paper, the main work is as follows:(1) this paper focuses on the research and analysis of the support vector machine algorithm-- the most representative algorithm in data mining, and gives the relevant extraction rules of the support vector machine algorithm and the related theoretical basis, To compared and analyzed the posterior probability support vector machine algorithm and the traditional data mining algorithms.(2)To construct the model based on the data mining technology, simulate by the using of the MATLAB platform, implemented the customer churn prediction based on posterior probability support vector machine mode.(3)Based on the improved model, using two kinds of data-the data of Mobile Corporation customer and the data of UCI(University of CaliforniaIrvine)to simulate, and the simulation results show that to find the main characteristics of the loss customers will help telecommunications company take effective measures to improve customer loyalty as soon as possible.(4) To introduce the complex networks, analyzed and compared the traditional model of infectious diseases, put forward an improved SKIR model on complex network application, it verified that the improved model has clear description of the information spreading on complex networks, and have good suppression effect to the spread of rumors.