一個企業內網路使用記錄之資料發掘架構

A Data Mining Framework For Intranet Logs

以往對全球資訊網的資料發掘主要侷限於網際網路上，而發掘的規則也分別是有關網頁內容或使用者紀錄，沒有看到兼具兩者的知識。然而，隨著企業內網路的運用日益廣泛，有關其資料發掘的研究也日益重要。本研究把「資料內容」與「使用者紀錄」加以結合，另方面，也考慮到發掘之前置處理工作，提出一個具有學術價值及實務可行的使用紀錄資料發掘架構，來發掘企業內網路中所隱含的兼具「特性規則」及「關聯規則」之使用行為知識。同時以JAVA建置一個雛形系統，來驗證其可行性。

Most of past researchers have only paid attentions to data mining on Internet. Their mining rules are either related to Web content or Web Usage, but not to both. Since the applications on Intranet have become widespread, its related data mining is now important. This research has combined “Web content” and “Web usage” data, and proposed an intranet data mining framework, which should be academically valuable and practically feasible. It includes the pre-processing phase of data mining. The framework would be used to find out the combination of characteristic rules and association rules, which would imply some usage behaviors. In addition, we will implement a prototype in JAVA to test the feasibility of the framework.