網頁探勘的方法只要分成三類:網頁使用模式探勘、網頁結構探勘及網頁內文探勘。其中網頁使用模式探勘分成三個步驟:前置處理、模式發掘、模式分析。本論文則在前置處理中先建構SPTN網頁架構模型,並利用此STPN網頁架構模型來填補網站日誌檔中的使用者瀏覽路徑。在模式發掘中,統計使用者連結網頁次數與瀏覽時間,利用馬可夫矩陣求得其極限機率,與每個網頁的瀏覽率與所佔的時間比例。最後藉由我們所預估的值與實際從網站日誌檔中統計的資料做比較,以提供網站管理者可適時地調整網頁架構或網頁內容,使得使用者瀏覽更方便,且達到網站負載之最大效益。 The approaches of web mining have been classified into three types, web usage mining, web structure mining, and web content mining. The web usage mining consists of three phases: data preprocessing, pattern discovery, and pattern analysis. In this paper, we use STPN web structure model to solve path completion and use the link graph to calculate a transition matrix containing one-step transition probabilities between the states in the Markov model. Finally, we compare the probability of linking each page by using Markov model and the probability that we predict. We provide the administrators adjusting the web structure and content to make the navigation of user more convenient.