Research On The Mining Of Potential Web Customers

With the rapid development of internet and communication technology and the gradual maturity of the market, the world economy has currently entered a global, electronic era. Various products and services among enterprises are smaller and narrower so that the traditional market strategy which takes the production as the center and selling as the purpose has gradually been replaced by the new trend, customer-and-service-centric strategy. Consumers are increasingly dependent on the network platform for commodity trading; Business elites have recognized that enhancing good relationships with customers is the crucial element to success in this electronic commercial age. That means whoever is able to grasp the customer demand trends, strengthen the relationships with customer, and effectively develop identify and manage customer resources will win this increasingly intense competition.Therefore, potential customer exploitation will provide accurate reference information and scientific data basis for the business to adjust its customer service strategy and effectively make decisions. Customer resources have naturally become the competitive focus, thus digging out the customers'demand trends and provide them with the personalized service come to the target of enterprises.Meanwhile, as people need to extract the information and knowledge from the huge amounts of data urgently, data mining technology in various industrial applications are becoming increasingly fierce. This article is trying to obtain the specific applications of data exploring techniques to the websites customer management to facilitate decision-making people to develop better strategies; at the same time, taking the exploration of potential users of the book sale websites for example, this paper is going to investigate the personalized service processes guided by the ideology of the data exploration.According to the principle of data mining algorithm, this paper will discuss and research the books sale websites, targeting to dig out potential customers, provide them with personalized service and expand the number of the customers. This research mainly includes aspects of information acquisition, data preprocessing, as well as decision-making in terms of potential users and etc. The followings are the primary tasks:First, using the user access paths to collect sample information of the new website customers; building the potential feature category after extracting their common features; summarizing the characteristics of the samples according to the new customers'feature category; making the classifications by Bayesian.Second, as for the unidentified customer information, taking its traits, with the help of the statistical decision tree, drawing its critical attribute, analyzing the key attributes of the sample and then adding those whose features are meeting the conditions to the existing categories and those not to new categories.Finally, the textbook website background, design and implementation of a statistical decision tree based on Bayesian filtering and potential users of excavation, the site of the textbook analysis of the database, and tap the potential users. Experiments show that the method is effective.