以信任為基礎之個人產出服務推薦系統

Abstract

個人產出服務(Peer Production Services)逐漸地將傳統的資產密集生產模式轉變成重度依賴資訊創造與分享的模式。越來越多的線上使用者依賴此類服務，例如：新聞、文章、書籤，以及眾多分散於全球資訊網上的使用者產出內容(User-generated content)。然而，這些使用者產出的品質以及可信度並未有效的管理，如果沒有適當的機制來衡量使用者產出的品質，結果將導致資訊過載(Information overload)。本研究提出一個基於社會網路信任的推薦系統，藉由信任運算，使用者產出的品質與可信度得以適當的衡量。本研究並整合了兩種著名的模糊邏輯應用-「模糊推論系統」以及「模糊多準則決策方法」用以支援服務選擇的決策制定。實驗結果顯示，本研 究提出的系統能夠有效的提升使用者產出服務的品質，進而克服資訊過載的問題。最後，本研究建置了一個以信任為基礎的「個人產出新聞系統(Social news system)」用以呈現系統的可能性應用。

Peer Production, a new mode of production, is gradually shifting the traditional, capital-intensive wealth production to a model which heavily depends on information creating and sharing. More and more online users are relying on this type of services, such as news, articles, bookmarks, and various user-generated contents, around World Wide Web. However, the quality and the veracity of peers’ contributions are not well managed. Without a practical means to assess the quality of peer production services, the consequence is information-overloading. In this study, we present a recommender system based on the trust of social networks. Through the trust computing, the quality and the veracity of peer production services can be appropriately assessed. Two prominent fuzzy logic applications - fuzzy inference system and fuzzy MCDM method are utilized to support the decision of service choice. The experimental results showed that the proposed recommender system can significantly enhance the quality of peer production services and furthermore overcome the information overload problems. In addition, a trust-based social news system is built to demonstrate the application of the proposed system.