面對競爭激烈的時代，企業無不採取電腦化來提高效益。利用資料挖掘技術，可以幫助企業擷取龐大資料庫裡隱藏的有用資訊。資料挖掘是現今相當熱門的研究領域，其中循序式樣(Sequential Pattern)的挖掘是從循序的資料庫中找出經常出現的循序式樣。這個技術已經受到廣泛的應用，大致可區分為消費者購買行為分析、網頁點閱行為分析，以及時間趨勢分析等。而在探討消費者購買行為的時序性上，大多藉著候選序列的產生及驗證，以漸進地過程來產生常見序列型樣。近幾年更有學者提出封閉性偏序關係(Closed Partial Order)來取代循序式樣的方法。封閉性偏序關係可以獲得精確且有趣的順序資訊。儘管封閉性偏序關係可以明顯表示出項目間的順序關係，但卻無法得知項目間的詳細時間資訊。因此本研究著重於研究由時間區間循序資料庫(Time-Interval Sequence Databases)中找出包含時間資訊的封閉性偏序關係(Closed Time-Interval Partial Orders)，並發展有效率的演算法從循序資料庫中找到包含時間資訊的封閉性偏序關係。 Facing competitively intense era, the enterprise adopts the computer to enhance the benefit all. Using data mining technique can help enterprise to mine useful information in huge databases. Data mining is a popular domain of research in resent years, and sequential pattern mining is one of the popular techniques. Sequential pattern mining is used to disclose frequent sequential patterns in databases. This technique has broad applications, including marketing and behavior analysis of customer purchasing, web browsing, and time series analysis and so on. However, researches about the sequence of consumers’ shopping behaviors are mostly in producing frequent sequential patterns by generating and verifying candidates of sequential patterns. Recent years, some scholars proposed closed partial order method in place of finding sequential patterns, which can find accurate and interesting information in sequence database. Nevertheless, the closed partial order can display the ordering relationship of items, but it can’t discover the detailed information of times. In this paper, our research focuses on finding closed time-interval partial orders in item-interval sequence databases. In addition, we develop an efficient algorithm to reveal closed time-interval partial orders.
