臺灣博碩士論文加值系統

查詢擴展(Query Expansion)主要是被設計用來克服使用者在搜尋過程中，送出過短查詢字的問題，並且已經被應用於多個商業產品上。透過語意的計算，這方法試著分析使用者實際的資訊需求，並依此來進行擴展的行為。在於資訊過載與有效降低查詢使用門檻的問題上，查詢擴展或許是一項重要的關鍵技術。再來是在於目前檢索系統嚴重缺乏使用者端的考量，結果直接造成檢索效能低落的窘境。基於這些問題上，我們提出一模式，透過線性最小平方配適法(Linear Least Squares Fit)來分析個別使用者的歷史搜尋過程，並以矩陣的形式表現出個人化設定檔。這模型可以依據不同使用者背景調適出最適當的擴展查詢字詞。針對於檢索精確度上的提升，我們同樣也考慮了重新權重查詢字(Query Words Re-weighting)與文件集合(Document Pooling)。最後在檢索出一群與查詢字詞最相關的文章集合後，我們再度提出考量進四個影響排序結果因素的Rscore排序法，並利用此法將最相關的文件呈現到使用者面前。最終我們透過實驗設計實作並評估我們的演算法，結果顯示我們的方法不僅效能佳且達到個人化。

Query Expansion was designed to overcome the barren query words issued by the user and has been applied in many commercial products. This treatment tries to expand query words to identify users’ real requirement based on semantic computation. It may be critical to deal with the problem of information overloading and diminish the using threshold, however the modern retrieval systems usually lack user modeling and are not adaptive to individual users, resulting in inherently non-optimal retrieval performance. In this study, we propose the LLSF method based on each individual search history to automatically generate specific personalized profile matrix. By which to generate context-based expanded query words. Considering the accuracy of retrieving performance, we process query words re-weighting and document pooling algorithm to achieve this goal. Finally, the documents list is ranked by the way of stressed density distribution modeling. And the experimental results show that our framework corresponds to personalization and the performance is very promising.

中文摘要 i
英文摘要 ii
誌 謝 iii
Contents iv
Tables vi
Figure Illustrations vii
Chapter 1 Introduction 1
1.1 Research Background and Motivation 1
1.2 Research Objective 2
1.3 Research Contribution 2
1.4 Research Restriction 3
2.1 Query Processing 4
2.2 Document Retrieval 4
2.3 Query Expansion 4
2.3.1 Review of Refining Short Query 4
2.3.2 Categories of Query Expansion 5
2.4 Personalization 6
2.4.1 Personalization Achievement 6
2.4.2 Personal Data Construction 6
2.5 Clustering Analysis 7
2.5.1 Basic Concepts 7
2.5.2 Clustering method 8
2.6 Document Pooling 8
3.1 Introduction 10
3.2 Retrieval Method 11
3.2.1 Term Weighting 12
3.2.2 Retrieving Model 13
3.2.3 Retrieving All Possible Result 14
3.2.4 Ranking Result 15
3.3 Query Expansion 19
3.3.1 Probabilistic Models of Query Expansion 19
3.3.2 LLSF Models of Query Expansion 20
Chapter4 Experiment Design 28
4.1 System Design 28
4.2 Experimental Data Sets 29
4.3 Sub-Component Description 30
4.3.1 Word Recognition 30
4.3.2 Search Component 32
4.3.3 Query Expansion Component 33
4.3.4 Rscore Ranking 36
4.3.5 System Interface 37
Chapter 5 Evaluation of PNQES 39
5.1 Evaluation Method 39
5.1.1 Experimental Subject 39
5.1.2 Evaluation Variable 39
5.1.3 Evaluation Procedure 39
5.1.2 Retrieval and Ranking Statistics 41
5.2 Experiment Result 44
Chapter 6 Conclusion and Future Research 48
Reference 50
Appendix A: EVAL Scored Card A and B 53
Appendix B: Definition of POS 60

Croft, W. B., & Harper, D. J. (1979). Using probabilistic models of document retrieval without relevance information. Journal of Documentation, 35(4), 285 - 295.
Deerwester, Fumas, Landauer, & Harshman. (1990). Indexing by intent semantic analysis. Paper presented at the JASIS.
Dunham, M. H. (2003). Data mining introductory and advanced topics. Upper Saddle River, New Jersey: Prentice Hall.
Eirinaki, M., & Vazirgiannis, M. (2003, February). Web mining for web personalization. Paper presented at the ACM Transactions on Internet Technology (TOIT).
Gauch, S., Chafee, J., & Pretschner, A. (2004). Ontologybased personalized search and browsing. Web Intelligence and Agent Systems, 1(3-4), 219-234.
Glen Jeh, & Widom, J. (2003, May). Scaling personalized web search. Paper presented at the Proceedings of the 12th international conference on World Wide Web.
Harman, D. (1992 June). Relevance feedback revisited. Paper presented at the Proceedings of the 15th annual international ACM SIGIR conference on Research and development in information retrieval, New York.
HE, H.-z., HE, P.-l., GAO, J.-f., & HUANG, C.-n. (2002). Query expansion based on the context in chinese information retrieval. Joumal of Chinese Infomation Processing, 16(6), 32-37.
Hofmann. (1999). Prohobilistic latent seniontic indexing. Paper presented at the SIGIR.
Jackson, P., & Moulinier, I. (2002). Nature language processing for online application text retrieval, extraction and categorization. Amsterdam/Philadelphia: John Benjamins Publishing Company.
Jansen, B. J., Spink, A., & Saracevic, T. (2000,March 1). Real life, real users, and real needs: A study and analysis of user queries on the web. Information Processing & Management, 36(2), 207-227.
K. F. Jea and P. Y. Hsu, “An Entropy-based Ranking Method for Internet Search,”
Proceedings of the 2000 Workshop on Internet and Distributed
Systems, May 11-12, 2000, pp. 115-123.
Liu, F., Yu, C., & Meng, W. (2002, November 4--9). Personalized web search by mapping user queries to categories. Paper presented at the In Proceedings of CIKM, McLean, Virginia, USA.
Mandala, R., Tokuanga, T., & Tanaka, H. (1999). Combining mutiple evidence from different types of thesaurus for query expansion. Paper presented at the SIGIR.
Manning, C. D., & Schütze., H. (1999). Foundations of statistical natural language processing (Massachusetts Institute of Technology ed.). Cambridge: MIT Press.
Miller, G. A. e. a. (1990). Introduction to wordnet: An on-line lexical database. International Joumal of Lexicography, 3(4), 235-312.
Mobasher, B., Cooley, R., & Srivastava, J. (2000,August). Automatic personalization based on web usage mining. Communications of the ACM, 43(8), 142-151.
Qiu, Y. (1995). Automatic query expansion based on a similarity thesaurus. ETH Zurich.
Robertson, S. E., & Sparck Jones, K. (1976). Relevance weighting of search terms. Journal of the American Society for Information Sciences, 27(3), 129 - 146.
Rocchio. (1971). Relevance feedback information retrieval. In The smart retrieval system-experiments in automatic document processing (pp. 313 - 323). Kansas: Prentice-Hall.
Salton, G., & Buckley, C. (1990). Improving retrieval performance by relevance feedback. 41(4), 288 - 297.
Salton, G., & McGill, M. J. (1983). Introduction to modern information retrieval. New York: McGraw-Hill Book Company.
Shen, X., Tan, B., & Zhai, C. (2005, Augest 15-19). Contextsensitive information retrieval usingimplicit feedback. Paper presented at the SIGIR, Salvador,Brazil.
Speretta, M., & Gauch, S. (2004). Personalizing search based on user search history. Paper presented at the CIKM.
Srivastava, J., Cooley, R., Deshpande, M., & Tan, P.-N. (2000,Jan). Discovery and applications of usage patterns from web data. Paper presented at the SIGKDD Explorations Newsletter.
Sugiyama, K., Hatano, K., & Yoshikawa, M. (2004, May 17 - 22). Adaptive web search based on user profile constructed without any effort from users. Paper presented at the WWW, New York, New York, USA.
Teevan, J., Dumais, S. T., & Horvitz, E. (2005, August 15–19). Personalizing search via automated analysis of interests and activities. Paper presented at the SIGIR, Salvador, Brazil.
Witten, I. H., & Frank., E. (2000). Data mining: Practical machine learing tools and techniques with java implementations. San Diego: Morgan Kaufmann.
Yang, Y., J. Carbonell, R. Brown, T. Pierce, B.T. Archibald and X. Liu, 1999, "Learning approaches for detecting and tracking news events", IEEE Intelligent
Systems, Vol. 14, No. 4, pp. 32-43.
ZHANG, M., SONG, R., LIN, C., MA, S., JIANG, Z., LIU, Y., et al. (2002). Expansion-based technologies in finding relevant and new information. Paper presented at the TERC.