Design and Implementation of Access Mechanism and Data Mining in Electronic Commerce

在此一計畫中，我們設計並完成了電子 商務系統之使用機制，並對於以網路交易 資料為主的商品相關性與消費者行為的相 關研究。電子商務是指消費者可以透過網 際網路來進行消費的網路交易機制。根據 文獻，以往提出的資訊搜尋方法，皆是分 析傳統交易方式下所產生的交易資料。我 們發現此類型的搜尋方法難以分析電子商 務服務所衍生的新型態客戶交易行為，即 客戶在網路上的交易路徑，交易網站與交 易商品的相關性。因此，我們在此一研究 計畫中針對此問題加以深入研究，並提出 一個架構來模擬現有的電子商務使用機制 以產生完整的消費者交易資料，並設計一 系列的演算法來分析此交易資料來對商品 相關性與消費者行為進行資訊搜尋。In this report, we explore a new data mining capability which involves mining Web-based association rules generated from electronic commerce services in a distributed commodities-provided environment where customers can seek for commodities of interest, travel from one node (i.e., Web page, whose location is unique in the Internet) to another node via the corresponding facilities (i.e., hyperlinks) provided, and make transactions immediately. It is important to note that, since customers are traveling the commodities-providing Web sites to search for desired commodities and make transactions in the corresponding Web nodes, Web traversal sequences and transaction records are both utilized to mine the Webbased association rule and customer behavior, showing that this problem is very different from the traversal pattern problem and others which are mainly based on customer transaction items