| With the rapid development of the Internet, web-based online bookstores have drawn much public attention. The change from bookstores to online bookstores reduces both labor and sales cost for operators, and it attracts customers from different regions. However, it has weakened the interaction between the users and the websites, which is not convenient for vendors to recommend newly-published books to users and promote sales. At the same time, various categories of books and the users aimless browsing will greatly increase the difficulty to understand users’ preferences. In this way, it fast understands of users’ interests and recommendation of properly books that counts in the website book sales.Before the development of a personalized online bookstore recommendation system, we first need to consider the main functions of the site should be realized. Through the relevant survey we found that in the preparation of the online bookstore, registrants should be divided into two roles, naming members and management personnel. After logging in, you can view the detailed information and the option to purchase a variety of goods. After the administrator login, you can manage membership, merchandise and orders.The system intends to use content-based retrieval to recommend the best-selling book before N, and a variety of user interest books based on a user’s preference. When recommending the realization and the most user’s interesting N books, this algorithm can obtain statistical customer purchase history. If you buy a certain category of books most of these books then a user has a higher interest degrees.This thesis compares and analyzes e-commerce recommendation systems both in China and other foreign countries. And then, we analyze its relevant theory, methods, technology, evaluation as well as application of e-commerce recommendation system. It is a hybrid recommendation system, which uses several recommend skills, and is implemented with asp.net platform. |
PDF Full Text Request