臺灣博碩士論文加值系統

由於網際網路的盛行，許多傳統的商家看好網際網路所帶來的商機，紛紛建置了電子商店及購物網站。而這些網際網路上的購物網站，大部分只針對其本身所販售的商品提供瀏覽、查詢、比價以及訂購的服務，並無法根據使用者對於某些商品的特殊偏好提供多元化的商品資訊及購買建議。假設使用者想購買一組電腦配備，其必須一一造訪各個電腦購物網站，自行搜集、過濾並匯整所獲得之相關商品資訊，對於商品價格及付款方式等其他相關購物資訊皆需自行進行比較，對於使用者來說既費時又費力。正因如此，網際網路購物代理人便應運而生。
網際網路購物代理人是一種介於使用者與購物網站之間的中介者(或伺服器)，它不單能幫助使用者搜集、過濾及匯整購物商品的相關資訊，也能夠針對商品進行價格比較，甚至定期遞送相關的商品資訊給使用者參考，節省了使用者在購物上所花費的時間與金錢。網際網路購物代理人可以同時查詢多個購物網站的商品資料，並將所抓取回來的商品資料以一致的格式呈現給使用者，所提供給使用者的相關商品資訊較單一購物網站更完整。然而，過去有關於網際網路購物代理人的研究大都侷限於探討如何快速地將多個來源的購物網站資訊加以過濾、匯整以及價格比較，顯少會對於使用者偏好及個人化購物去做探討，以針對使用者的購物偏好做進一步的商品購買推薦與建議。由於網際網路購物代理人所能提供的資訊量遠較一般購物網站來得豐富而且繁雜，因此運用個人化技術來協助使用者簡化購物資訊，並且依據使用者個人的購物偏好及習性來做進一步的商品購買建議或最新的商品資訊即時傳遞，在網際網路購物代理人的研究中更形重要。
本論文的目的，除了探討過去相關的研究外，便是要運用個人化的技術來建構一個智慧型的網際網路購物代理人，以提供使用者多元化而且精準的購物資訊。本論文並提出一個新的個人化技術，此技術運用一個使用者偏好矩陣來記錄使用者對於各項商品的偏好，以及其購物時瀏覽、查詢及購買的行為。此法不僅運算快速，而且實驗證明其對於使用者偏好的辨識率亦相當高。此外，我們也利用此技術設計並實作了一個智慧型的個人化網際網路購物代理人 － Oh!Hot 以驗證此法的實用性。相信藉由此篇研究論文及 Oh!Hot 網際網路購物資訊站，能夠將使用者偏好及個人化技術帶入另一新的應用領域，並使網際網路購物代理人的發展能夠更趨於完美。

With the recent expansion of the Internet, most merchants start in build their own Internet shop. The Internet shops supply its merchandise to users. Users can browse detailed merchandise information, look up for some merchandise he wants or buy some merchandise in needs. Sometimes, users will receive the mail about merchandise information from Internet shop. However, most of the Internet shop provides only merchandise of its own. No connection to other Internet shops, no price comparison, no user shopping preference and no merchandise recommendation considered.
Internet Shopping Agent (ISA) is kind of autonomous software that assists in searching the Internet for merchandise information on behalf of a user. ISA helps user to collect, filter and integrate merchandise information from multiple Internet shops. Some ISAs also provide price comparison and deliver the newest merchandise information to users periodically. Besides, ISAs provide more complete merchandise information than Internet shops. However, previous researches about ISA are restricted in integrate related merchandise information from multiple Internet shops rapidly, or price comparison. Discussion about user''s preference and personalized shopping are neglected. Also, making merchandise recommendation according to user''s shopping preference is not considered. Since ISAs provide large amount of merchandise information than Internet shops, it is more important for ISA to help users simplify merchandise information. Hence, applying personalization technologies to record/learn user''s shopping preference and to make shopping recommendation is needed in researches of ISA.
In this paper, we propose an Intelligent-Personalized Internet Shopping Agent (I-PISA) that not only helps users to collect, filter and integrate merchandise information, but also perform price comparison and make merchandise recommendation according to shopping preference of users. I-PISA use a new data structure called User Preference Matrix (UPM) to keep track of users'' behaviors (including user registration, browsing, querying, and purchasing…etc). According to UPM, I-PISA will make merchandise recommendation and automatically deliver merchandise information that match user''s shopping preference periodically. UPM not only evaluates user''s preference efficiently, but also provide high accuracy in judgement of user''s preference. To evaluate UPM, we apply UPM to implement a intelligent personalized Internet shopping agent called Oh!Hot. Oh!Hot is towards to create a more clever environment of Internet shopping in electronic commerce.

Chapter 1 Introduction
1.1 Motivation and Objective
1.2 Introduction of Internet Shops
1.3 Introduction of Internet Shopping Agents
1.4 Organization of this thesis
Chapter 2 Previous Works
2.1 Survey of Internet Shopping Agents
2.1.1 BargainFinder
2.1.2 Jango
2.1.3 BargainBot
2.1.4 PersonaLogic
2.1.5 ShopBot
2.1.6 Firefly
2.1.7 AuctionBot
2.1.8 Kasbah
2.1.9 Comparison among existing ISAs
2.2 Survey of personalization technologies
Chapter 3 System Architecture of I-PISA
3.1 Architecture of I-PISA
3.1.1 Data Collection and Classification Module (DCCM)
3.1.2 Data Processing Module (DPM)
3.1.3 Data Representation Module (DRM)
3.2 Advantages of I-PISA
Chapter 4 Personalization Technique of I-PISA
4.1 Introduction of User (Consumer) Behaviors
4.2 Principle of User Preference Matrix (UPM)
4.2.1 Basic Idea
4.2.2 Judgement of user''s preference
4.2.3 Judgement of like-minded users group
4.3 Variations of UPM
4.4 Experiments and analysis
4.4.1 Experiment environment
4.4.2 Preference evaluation
4.4.3 Results analysis
Chapter 5 Implementation
5.1 System Environment
5.2 Techniques of Oh!Hot implementation
5.2.1 Implementation of DCCM
5.2.2 Implementation of DPM
5.2.3 Implementation of DRM
5.3 Miscellaneous of implementation
5.4 An execution example of Oh!Hot
Chapter 6 Conclusions and Future Works

[1] CDnow. http://www.cdnow.com
[2] Amazon. http://www.amazon.com
[3] 1-800-Flowers. http://www.1800flowers.com
[4] Fedex. http://www.fedex.com
[5] PersonaLogic. http://www.personalogic.com
[6] Firefly. http://www.firefly.com
[7] BargainFinder. http://www.ac.com/services/cstar/cstar_child/ecpocketbf_cn.htm
[8] Jango. http://www.jango.com
[9] Bargainbot. http://www.ece.curtin.edu.au/%7Esaounb/bargainbot/
[10] Robert B. Doorenbos, Oren Etzioni, and Daniel S. Weld. “A Scalable Comparison-Shopping Agent for the World-Wide Web.” Agents''97 Conference Proceedings, ACM, 1997.
[11] Philip S. Yu. “Data Mining and Personalization Technologies.” Proceedings Sixth International Conference on Database Systems for Advanced Applications, 1999.
[12] Dean, R. “Personalizing your web site.” http://builder.cnet.com/Business/
Personal/index.html, 1998.
[13] M. Pazzani, J. Muramatsu, D. Billsus. “Syskill & webert: Identifying interesting web sites.” AAAI conference, Portland, 1996.
[14] Upendra Shardanand. “Social Information Filtering for Music Recommendation.” M.S. thesis, Massachusetts Institute of Technology, 1994.
[15] Upendra Shardanand and Pattie Maes. “Social Information Filtering: Algorithms for Automating “Word of Mouth”.” Proceedings of the Computer-Human Interaction Conference (CHI''95), Denver, CO, 1995.
[16] Liren Chen and Katia Sycara. “WebMate: A Personal Agent for Browsing and Searching.” Agents ''98 Conference Proceedings, ACM. Pages 132-139, 1998.
[17] Marko Balabanovic and Yoav Shoham. “Learning Information Retrieval Agents: Experiments with Automated Web Browsing.” Proceedings of the AAAI Spring Symposium Series on Information Gathering from Heterogeneous, Distributed Environments, Pages 13-18, 1995.
[18] Marko Balabanovic and Yoav Shoham. “Fab: Content-Based, Collaborative Recommendation.” Communication of the ACM, Vol. 40, No. 3, 1997.
[19] P. Maes. “Agents that Reduce Work and informatioin Overload.” Communication of the ACM, Vol. 37, No. 7, pp. 31-40. July 1994.
[20] Guttman, R. H. and Maes, P. “Agent-mediated Integrative Negotiation for Retail Electronic Commerce”, 1998. http://ecommerce.media.mit.edu
[21] AuctionBot. http://auction.eecs.umich.edu
[22] Kasbah. http://kasbah.media.mit.edu
[23] Paolo Atzeni and Giansalvatore Mecca. “Cut and Paste.” Proceedings of the 6th ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, pp. 144-153. Tucson, Arizona. ACM Press 1997.
[24] Paolo Atzeni, Giansalvatore Mecca and Paolo Merialdo. “To Weave the Web.” Proceedings of 23rd International Conference on Very Large Data Bases (VLDB''97), pp. 206-215. August 25-29, 1997, Athens, Greece.
[25] Salton G. “Introduction to Modern Information Retrieval.” Addison-Wesley Publishing Company.
[26] Bassam Aoun. “Robots and the Web: An Electronic Shopping Agent.” AUSWEB 96’, 1996.
[27] L. Ardissono, C. Barbero, A. Goy and G. Petrone. “An Agent Architecture for Personalized Web Stores.” Agents ’99 Conference Proceedings, ACM. Pages 182-189, 1999.
[28] Robert Inder, Matthew Hurst and Toshikazu Kato. “A prototype agent to assist shoppers.” Computer Networks and ISDN Systems, Vol. 30, Pages 643-645. 1998.
[29] Jin-gang Kim, and Eun-seok Lee. “Intelligent Information Recommend System on the Internet.” IEEE International Workshops on Parallel Processing proceeding, 1999.
[30] Ravi Kumar, Prabhakar Raghavan, Sridhar Rajagopalan, and Andrew Tomkins. “Recommendation systems: a probabilistic analysis.” 39th Annual Symposium on Foundations of Computer Science Proceeding, 1998.
[31] Del I. Hawkins, Roger J. Best, and Kenneth A. Coney. “Consumer Behavior: Implications for Marketing Strategy,” Fifth Edition. Richard D. Irwin, INC, 1992.
[32] 簡貞玉譯. “消費者行為學.” 國立編譯館主譯, 五南圖書出版公司印行, 1995.
[33] 熊祥林著. “消費者的知覺.” 理明叢書, 消費者行為序列, 1990.
[34] 蔡瑞宇編著. “顧客行為學.” 天一圖書公司發行, 1996.
[35] 佐藤信著,官感檢查研究小組譯. “統計官感檢查法.” 先鋒企業管理發展中心發行, 1988.
[36] 蕭鏡堂編著. “市場研究(Marketing Research)－理論、方法、運用,” 第二版. 興利企管顧問公司, 1991.