知識社群系統之建構-以交大圖書館為例

Abstract

虛擬社群是最近新興的議題，具有類似興趣的使用者在網路上漸漸形成群組，利用網路分享彼此的知識，能大幅增加資訊的再使用率和資訊潛在價值，進而達到共同學習的效果。知識社群最能發揮內隱知識的傳遞及知識的創新，為知識管理社會化之工具。因此，本研究利用知識社群的理念，開發一個具個人化推薦功能的知識社群系統，將知識社群的理念應用至圖書館的環境中，讓圖書館的使用者能藉由與同好間的交流互動而學習知識。其中的推薦功能是利用關聯規則分析及協力式過濾，根據使用者過去的借閱記錄，自動尋找和其興趣類似的同好，再依同好的借閱記錄來推薦使用者圖書分類及社群內的討論區，可以減少使用者找尋並加入社群所花費的時間。本研究的目的在於：(1)設計開發一個圖書館知識社群系統；(2)經由知識社群提供個人化推薦服務以加快使用者搜尋速度；(3)加強使用者、資源及專家之間的互動，提高使用者的學習效率。在使用者試用過雛形系統後，讓使用者填寫需求及滿意度問卷，調查結果顯示使用者需要且滿意這種知識分享模式，而個人化推薦功能也具有相當的參考價值。因此本研究所採用的知識社群模式可作為日後企業或學校建立知識管理系統之參考。

Virtual community is a new topic in knowlodge management. Users with same interests may form a community on internet. By sharing and interchange information with others on internet, users are able to learn from peers and thus enhance the reusability and potential value of information. The learning community is a good tool for knowledge socialization. We can transfer implicit knowledge to others and creat new knowledge easily in knowledge community. In this paper we adopt this concept and implement a knowlodge commuity system which can recommend user book’s classification and relate discussion forum automatically on National Chiao Tung University (NCTU) library. The recommendation method combines fuzzy apriori algorithm with collaborative filtering method. First, we use fuzzy apriori algorithm to discover user group which have the same preference for collaborative filtering. Second, we use collaborative filtering to find a recommend list which contains books collecting from other patrons’ borrowing history. Finally, we use combine recommend list with discussion forum classification to make recommendation. The purpose of this paper are: (1) implementing a knowledge community system in library; (2) providing personalize recommendation service to reduce user searching time; (3) enhancing the interactions among the users-resources, users-users and usersexperts to increase the learning effectiveness. After using the knowledge community prototype system, we send a questionary to user. The results show that users not only need but also be satisfied with the functions of knowledge community system.