電子商務中促進個人化推薦之服務中介者探討

隨著網際網路的發展及資訊科技的進步，網路逐漸成為現代人接受資訊十分重要的管道之一，其低成本且分散式的特性，使得越來越多個人或企業透過網路散播資訊或提供商業服務，然而一旦資訊過多且雜亂，對網路使用者來說反而造成負擔，這就是所謂的「資訊超載」問題。推薦系統根據使用者的喜好，幫助使用者從大量的資料中篩選出他所需要的資訊，商業網站經常透過此種系統的運作，推薦顧客喜歡的商品或服務給他們，以增加交易成功的機會、提高顧客對網站的忠誠度。推薦系統常用的技術主要可歸納為兩種，合作式推薦與內容導向式推薦，這兩種技術各有所長，相對的也各有其限制，同時結合兩種技術的混合式推薦才能截長補短，獲得較好的推薦效果；而目前關於使用者的偏好分散保存於不同網站裡的架構較適有利於合作式推薦的使用。本篇論文為此提出了服務中介者的架構，能夠集中保存使用者對不同服務項目的偏好，在不直接透露使用者偏好的情況下，服務提供者亦能從我們的中介者取得推薦的參考，提升推薦的品質；本研究著眼於網際網路上提供資訊商品的電子商務網站，並且以貝式分類法為基礎，計算推薦的結果；最後以實驗的方式說明架構的可行性，結果顯示集中保存使用者偏好的方式確實對推薦的效果有助益，也能提高資訊的使用率。

關鍵字

推薦系統；內容導向式推薦；貝式分類法；電子商務；個人化服務

並列摘要

Advancement of Internet and information technology brings some phenomenons: a. Internet becomes one of our important channels to access new information; b. more and more people and enterprises spread information or provide services through the Internet. However, the situation often causes “Information Overload problem,” which means too much information to be produced and users have too little time to properly digest it. Fortunately, Recomender Systems are one of the well-known solutions to information overload by helping users filter out unnecessary information. Such systems are thus in widespread use by service providers to enhance E-commerce sales. Based on how recommendations are made, recommender systems are usually classified into two main categories: collaborative recommendations and content-based recommendations. Both two techniques have their advantages and limitations, and it is proved that combining the two methods have more satisfied recommendation results. However, current architecture which user profiles are decentrally stored in different websites facilitates collaborative recommendations instead of content-based methods. In this thesis, we propose a service mediator system which mediates between the users and the service providers to centrally collect user profiles and without directly releasing user profiles, service providers would gain the “recommendation references” from the serive mediator. We focus on information goods in E-commerce, and the recommendation technique is based on Naïve Bayes Classifier. Finally we apply several experiments to test the feasibility of our system.

並列關鍵字

Recommender System ； Contnet-based Recommendation ； Naive Bayes Classifier ； E-Commerce ； Personalization

參考文獻

[1] A.I. Schein, A. Popescul, L.H. Ungar, and D.M. Pennock, “Methods and Metrics for Cold-Start Recommendations,” Proc. 25th Ann. Int’l ACM SIGIR Conf., 2002.

[3] Adomavicius, G., Tuzhilin, A., “Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions,” IEEE Transactions on Knowledge and Data Engineering, vol. 17, No.6, 2005.

[4] Apte’, C., Damerau, F., Weiss, S. M., “Automated Learning of Decision Rules for Text Categorization,” ACM Transaction on Information Systems, vol.12, No.3, pp.233-251, 1994.

[5] Balabanovic, M., Shoham, Y., “Fab: Content-based, Collaborative Recommendation,” Comm. ACM, vol. 40, no.3, pp.66-72, 1997.

[6] Berkovsky, S., “Ubiquitous User Modeling in Recommender Systems,” in Proc. of the UM Conference, Edinbirgh, UK, 2005.

國際替代計量

電子商務中促進個人化推薦之服務中介者探討

全文下載

主題瀏覽