Design and Implementation of an Adaptive Web Warehousing System

在網際網路成為新一代媒體的今日， 如何從龐大且雜亂無章的瀏覽記錄挖掘出 有價值的使用者行為，便成為所有網路服 務者亟需面對的課題。本研究之主要目的 為建置Web 環境應用所需的資料倉儲，包 含資料轉換程序與資料倉儲系統架設，最 後整合資訊勘測的諸項技術，包括相關 性、分類與瀏覽路徑等，以尋求隱含於大 量使用者資訊中的行為模式，此勘測結果 可提供使用者個人化的資訊服務，亦可回 授至本倉儲系統以動態的調校資料儲存結 構，維持系統最佳化，此一具可適性之特 點將特別適用於目前線上的Web 環境。具 體而言，整合以上各項技術成為一適合 Web 的應用環境且具線上分析勘測功能之 資料倉儲系統是本計畫的首要目標，因 此，我們針對資料倉儲系統之需求而研發 資料擷取、儲存與勘測機制及其系統架構。As Internet technologies develop rapidly these years, more and more people are attracted by various web applications. How to collect and save the user data efficiently and to find the valuable knowledge from the huge data have become important topics. In this project, the high-efficiency data collection mechanism and corresponding data processing steps for web environments are first being developed. Second, the OLAP techniques are utilized to build a well-found Web warehousing system. Additionally, data mining capabilities, i.e., association, classification and traversal pattern, are devised to explore the user behaviors hidden in large amount of user data. The mining results offer personal services for Web users, and can be feedback for our system to trigger the dynamic adjustment of storage architectures for optimizing the system. This adaptive characteristic is most suitable for online Web systems. In concrete, integrating the technologies mentioned above to build an adaptive web warehousing system with the OLAM functionality is our objective of this project