本論文設計和實作了一個新的百科知識編輯系統─NUPedia,相較於以往如Wikipedia等線上百科以多人匿名協作單一知識條目的方式,NUPedia提供了一個平台讓所有人都可以編輯、協作、及管理屬於自己的知識。本系統也提供了多媒體內容編輯工具、條目協作者、版本控制、條目分支版本、推薦機制、社群分享等功能,讓知識的編輯、協作與傳播更加的容易。 另外,本論文也針對NUPedia的推薦功能提出了關聯詞彙推薦機制。NUPedia可以透過查詢關聯詞彙資料庫,來推薦相關的條目給使用者。其中,本論文透過對大量新聞文章做斷詞與關鍵字處理,再對詞彙間的co-occurrence作統計,最後得到關聯詞彙資料庫並應用於上述之推薦系統中。
In this thesis we designed and implemented a new on-line encyclopedia system called NUPedia. Unlike previous encyclopedia system such as Wikipedia, which provides only one version of a knowledge term co-edited by multiple anonymous authors, NUPedia provides a platform for authors to edit/co-edit and manage their own knowledge. The system also provides some useful features, including multimedia editing tool, knowledge collaborators, version control, knowledge forking, recommendation mechanism, and so on. For the recommendation mechanism, this thesis demonstrates a method to build a related-term database by calculating the co-occurrence of the key-terms in a huge amount of news articles. NUPedia uses related-term database to recommend related knowledge to the users.