| Summary: | 碩士 === 中華大學 === 資訊管理學系(所) === 96 === The purpose of this research is to improve marketing efficiency of online bookstore by using data mining. Its principle is to apply association analysis and cluster analysis to produce personalized recommending book list for each customer to achieve one-to-one customizing service, and to increase the response rate or purchase rate. There were two kinds of data sources in this study: database data mining and web mining. In web mining, to determine the characteristics of customers, the number of web pages appeared the name of potential customers and the keywords of selling books are calculated by search engine from the Internet. Moreover, we used association analysis and cluster analysis to form association model and cluster model to build the community of customers and books. The e-marketing can be completed without customer database. Compared to conventional marketing response rate, which is only about 0.1~1%, the accuracy rate of web mining is up to 40-50%, and this is sufficient enough to increase new customers for online bookstores.
|
|---|
