基於模糊權重資訊檢索整合技術之推薦系統

Recommender System Based on Integrated Technique of Fuzzy Weight and Information Retrieval

10.6188/JEB.2008.10(3).12

鄭景俗(Ching-Hsue Cheng)；陳智賢(Jr-Shian Chen)；蘇勇戩(Yung-Chien Su)

模糊OWA運算子 ； 推薦系統 ； 資訊檢索 ； 模糊查詢 ； Fuzzy Order Weighted Averaging Operator ； Recommender System ； Information Retrieval ； Fuzzy Query

電子商務學報

10卷3期（2008 / 09 / 01）

781 - 803

繁體中文

面對資訊超載的時代，資訊檢索能有效率地提供更有價值的資訊也能減輕資訊超載的問題。資訊檢索的技術被廣泛應用於各種領域，例如：搜索引擎、資料探勘等，而推薦系統便是其中一例。近年來，電子商務蓬勃發展，推薦系統也被應用在增加消費者滿意度以及提高顧客忠誠度。但是推薦系統仍存在著一些問題有待克服。例如：協同過濾式的推薦方式，有「資料稀疏性」的問題，在新產品推出時或是沒有相關社群時會因為資料量太少而無法推薦；而內容導向式的推薦方式，會因為使用者無法明確提供興趣目標的描述時，而影響到相似物件的比對，造成找不到推薦產品的窘境。 本研究應用模糊權重資訊檢索整合技術來強化傳統的推薦系統，使用者對產品屬性不需有專業知識，僅需利用語言表達其對產品屬性的重要性，即可以查詢出所要的理想產品。本研究提供三種屬性權重運算方式：模糊OWA (order weighted averaging)、正規化模糊權重以及有經驗者偏好權重。使用者可依照個人不同情境的偏好，彈性地調整各項屬性權重分配，讓推薦結果更合理，達到有效推薦的目的。 本研究根據所提出的模式建置一個數位相機推薦的雛形系統，經由雛形系統的實驗可以得知：(1)模糊OWA權重運用於產品推薦上具有相當好的成效，(2)正規化模糊權重提供使用者較大彈性的查詢條件，正規化處理後的屬性權重更能客觀的反應使用者對屬性的重視程度，(3)有經驗者偏好權重必須透過大量問卷分析取得有經驗者對於各項屬性的重視程度，讓使用者能參考「群組意見」，作為購買決策依據，在推薦效果上相當有幫助。在三種權重運算方式的比較，模糊OWA權重運算方式所得到的推薦產品，皆能涵蓋另兩種權重運算方式所推薦的產品，具有較高成效性的推薦效果。

In the era of information overload, information retrieval technologies can solve the problem of information overload and extract valuable information efficiently from database. Information retrieval technologies are widely applied in many information technology fields such as: search engine, data mining, especially in Electronic Commerce recommender systems. In recent years, by the booming development in the Electronic Commerce, recommender system design has been utilized in Electronic Commerce websites to improve customer satisfaction and loyalty. However, there are two problems which have not been completely solved in traditional recommender systems: (1) the collaborative filtering recommendation approaches will be non-functional, because there is insufficient information in new products or related community. (2) The content-based recommendation approaches will be useless when users can not precisely point out their interesting and needs for products. This paper proposes a recommender system based on integrated technique of fuzzy weight and information retrieval to enhance the traditional recommender systems. By using linguistic expressions to define good attributes, the proposed system can make users find out their target goods without the priority knowledge of these goods. There are three weight operators provided in the proposed system: (1) fuzzy OWA operator, (2) fuzzy normalization operator, and (3) preference operator based on experienced users. With these weight operators, users can adjust the attribute weights more flexibly to make the searching results more reasonable to users, and, therefore, make the proposed recommender system more effectively in recommending products. Based on the proposed method, we developed a web-based prototype of digital camera recommender system. From the verification results for the prototype, there are three findings provided: (1) The fuzzy OWA operators perform very proper recommender results for users; (2) The fuzzy normalized weight operators offer more flexible query conditions for users, and represent the user concern for good attributes more impersonal. (3) The preference operator based on experienced users wills helpful recommender results by using the suggestion of the community. However, from the comparison result of three different operators, it indicates that recommender results, recommended from the fuzzy OWA weight operator, comprise the outputs from the other two operators, and, therefore, the proposed system provides better recommended results with higher coverage.

1. Balabanovic`, M.,Shoham, Y.(1997).Fab: Content-based, collaborative recommendation.Communications of the ACM,40(3),66-72.
2. Cheng, C.H.,Chang, J.R.(2006).MCDM aggregation model by ME-OWA and MEOWGA operators.International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems,14(4),421-443.
3. Cheng, C.H.,Chang, J.R.,Yeh, C.A.(2006).Entropy-based and trapezoid fuzzificationbased fuzzy time series approaches for forecasting IT project cost.Technological Forecasting and Social Change,73(5),524-542.
4. Cheng, P. J.,Yang, W. P.(1999).A new content-based access method for video databases.Information Sciences,118(1-4),37-73.
5. Filev, D.,Yager, R. R.(1998).On the issue of obtaining OWA operator weights.Fuzzy sets and Systems,94(2),157-169.
6. Fuller, R.,Majlender, P.(2001).An analytic approach for obtaining maximal entropy OWA operator weights.Fuzzy sets and Systems,124(1),53-57.
7. Goldberg, D.,Nichols, D.,Oki, B. M.,Terry, D.(1992).Using collaborative filtering to weave an information tapestry.Communications of the ACM,35(12),61-70.
8. Herlocker, J. L.,Konstan, J. A.,Terveen, L. G.,Riedl, J. T.(2004).Evaluating collaborative filtering recommender systems.ACM Transactions on Information Systems,22(1),5-53.
9. Hill, W.,Stead, L.,Rosenstein, M.,Furnas, G. W.(1995).Proceedings of the CHI`95.New York:ACM Press.
10. Kohrs, A.,Merialdo, B.(2001).Creating user-adapted websites by the use of collaborative filtering.Interacting with Computers,13(6),695-716.
11. Kuo, Y. F.,Chen, L. S.(2001).Personalization technology application to internet content provider.Expert Systems with Applications,21(4),203-215.
12. Lancaster, F. W.,Warner, A. J.(1993).Information Retrieval Today.Arlington, Virginia:Information Resources press.
13. Lee, D. S.,Kim, G. Y.,Choi, H. I.(2003).A web-based collaborative filtering system.Pattern Recognition,36(2),519-526.
14. Miller, B. N.,Albert, I.,Lam, S. K.,Konstan, J. A.,Riedl, J.(2003).MovieLens unplugged: Experiences with an occasionally connected recommender system.Proceedings of the IUI`03,263-266.
15. Miller, G. A.(1968).The magical number seven or minus two: Some limits on our capacity of processing information.Psychological Review,63,81-97.
16. Mooney, R. J.,Roy, L.(2000).Proceedings of the Digital Libraries.San Antonio, TX:
17. Resnick, P.,Varian, H. R.(1997).Recommender systems.Communications of the ACM,40(3),56-58.
18. Rustam, M. V.,Fei, J.(2005).A diversity-based method for infrequent purchase decision support in e-commerce.Electronic Commerce Research and Applications,4(2),143-158.
19. Schafer, J. B.,Konstan, J. A.,Riedl, J.(1999).Recommender systems in e-commerce.Proceedings of the ACM E-Commerce 1999 Conference,Denver Colorado:
20. Shardanand, U.,Maes, P.(1995).Proceedings of the CHI`95.New York:ACM Press.
21. Su, L.T.(2003).A comprehensive and systematic model of user evaluation of web search engines: I. theory and background.Journal of the American Society for Information Science and Technology,54(13),1175-1192.
22. Wang, J. W.,Cheng, C. H.,Chang, J. R.(2006).Flexible fuzzy OWA querying method for hemodialysis database.Soft Computing,10(11),1031-1042.
23. Yager, R. R.(1988).On ordered weighted averaging aggregation operators in multi-criteria decision making.IEEE Trans. Systems Man and Cybernetics,18(1),183-190.
24. Yager, R.R.(2003).Fuzzy logic methods in recommender systems.Fuzzy sets and systems,136(2),133-149.
25. Zadeh, L. A.(1965).Fuzzy sets.Information and Control,8(3),338-353.
26. Zadeh, L.A.(1975).The concepts of a linguistic variable and its application to approximate reasoning.Information Science,8,301-357.
27. 卜小蝶(1996)。圖書資訊系統檢索技術。台北:台灣學生書局。
28. 井上洋、天笠美知夫著、陳耀茂譯(1999)。模糊理論。台北:五南圖書出版公司。
29. 黃慕萱(1996)。資訊檢索中『相關概念』之研究。台北:台灣學生書局。