標題: | 應用社會性推薦於學術社群 Using Social Recommendation in Academic Community |
作者: | 粘怡祥 Yi-Hsiang Nien 柯皓仁 Hao-Ren Ke 資訊管理研究所 |
關鍵字: | 分群演算法;社會網路分析;推薦系統;機構典藏;學術社群;Clustering;Social Network Analysis;Recommendation System;Institutional Repository;Academic Community |
公開日期: | 2007 |
摘要: | 網際網路提供一個開放的平台,利用網路取得資訊已經成為最方便的管道。面對網路上充斥的大量資訊,使用者在找尋資訊時,相當的費時也不容易聚焦。推薦系統成為改善資訊過載問題的方法之一。使用者除了本身的主觀喜好之外,其行為容易受到人際關係的影響,於是虛擬社群與社會網路,成為許多使用者獲得資訊情報的最佳來源。
本研究主要的目在於提出結合主題概念萃取與社會網路分析之資訊推薦系統,以提供符合使用者需求之推薦資訊。本研究利用關鍵字分群演算法,萃取出使用者感興趣的主題概念;並且分析使用者社會網路,進行使用者分群,以形成主題社群;經由分析社群成員的主題偏好,預測使用者的潛在興趣,建構出更符合使用者需求的資訊推薦系統,以提升資訊推薦的品質。 Internet provides an open platform, which becomes a convenient channel to obtain information. Facing huge amount of information, users will spend a lot of time and become out of focus when they search information on Internet. Recommendation systems become one solution for information overloading. Besides the subjective preference of a user, interpersonal relationship will affect his/her behavior, and the concepts of virtual community and social network become one feasible information source for deriving interpersonal relationship. This thesis proposes a recommendation system combining topic concept extraction and social network analysis to meet users’ needs. This thesis uses the keyword clustering algorithm to extract topic concepts that users are interested in, followed by the formation of topic communities by analyzing the users’ social network to cluster the users. By analyzing the preferences of community members, the system can predict the potential interests and improve the quality of recommendation. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#GT009534510 http://hdl.handle.net/11536/39194 |
顯示於類別: | 畢業論文 |