整合資料倉儲與資料探勘於網站瀏覽分析

Abstract

隨著網路技術的成熟與網際網路的蓬勃發展，網路使用人數大幅增加，各行各業開始透過網站，將公司相關資訊迅速地提供給顧客，以期做好顧客關係管理，提高顧客滿意度。 經過時間的累積，網站伺服器上會留下大量的網站使用者瀏覽記錄。若能有效分析瀏覽記錄之歷史資料，對企業競爭力的提升，將有所助益。 本研究提出一套資訊系統架構，用來分析網頁使用記錄檔（Web Logs），整合資料探勘（Data Mining）技術，提供使用者群體推薦資訊，並利用資料倉儲（Data Warehousing）技術，提供經營者有效的決策資訊。 本研究架設了一個電腦週邊設備商模擬網站，用來驗證與展示本研究所提出的系統架構，最後再比較與討論本研究與其他相關文獻的差異。

Because of the highly mature network technologies and the explosive growth of Internet, the numbers of Internet users increase substantially. In order to keep good relationships with customers and raise customers’ satisfaction, more and more companies start to provide their customers easy and fast retrieval of business information and services through Internet, such as WWW. Consequently, Web logs on the Web servers keep growing as time passing by. The Web logs are not useless and not just wasting memory space. Instead, they are important sources from which we can retrieve useful information. Effective analysis of Web logs is helpful to promote enterprise competence. In this thesis, integrated system architecture for analyzing Web logs is proposed. It integrates Data mining technology to provide recommendations for Web users and Data Warehousing technology to provide decision information for managers. A test Web site simulating a company selling computer peripherals has been set up to verify the proposed architecture. The results are analyzed and presented in the thesis. Finally, some comparisons and discussions with other related papers are presented.