簡易檢索 / 詳目顯示
| 研究生： |
王榮聖 Wang, Rung Sheng |
|---|---|
| 論文名稱： |
非對稱性加權之排名學習機制 Leaning to rank with asymmetric discordant penalty |
| 指導教授： |
廖文宏
Liao, Wen Hung |
| 學位類別： |
碩士
Master |
| 系所名稱： |
理學院 - 資訊科學系 |
| 論文出版年： | 2009 |
| 畢業學年度： | 97 |
| 語文別： | 中文 |
| 論文頁數： | 75 |
| 中文關鍵詞： | 排名 、排名學習 、資料探勘 、非對稱加權 |
| 外文關鍵詞： | Information retrival, Asymmetric weight, RealRankBoost |
| 相關次數： | 點閱：134 下載：62 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
資訊發達的時代，資訊取得的方式與管道比起以前更方便而多元，但龐大資料量同時也造成了我們往往很難找到真正需要資料的問題，也因此資料的排名(ranking)問題就變得十分重要。本研究目的在於運用排名學習找出良好的排名，利用人對於某特定議題所給予的排名順序找出排名規則，並應用於資料探勘上，讓電腦可自動對資料做評分，產生正確的排序，將有助於資料的搜尋。
本研究分為兩部分，第一部份為排名演算法的設計，我們改良現有的排名方法(RankBoost)，設計出另一個新的演算法(RealRankBoost)，並且用LETOR benchmark實測，作為與其他方法的比較和效果提升的證明；第二部份為非對稱加權概念的提出，我們考量排名位置所造成的資料被檢視機率不同，而給予不同的權重，使排名結果能更貼近人類的角度。
With the innovation in computer technology, we have easier ways to access information. But the huge amount of data also makes it hard for us to find what we really want. This is why ranking is important to us. The central issues of many applications are ranking, such as document retrieval, expert finding, and anti spam. The objective of this thesis is to discover a good ranking function according to specific ranking order of the human perceptions. We employ the learning-to-rank approach to automatically score and generate ranking order that helps data searching.
This thesis is divided into two parts. Firstly, we design a new learning-to-rank algorithm named RealRankBoost based on an existing method (RankBoost). We investigate the efficacy of the proposed method by performing comparative analysis using the LETOR benchmark. Secondly, we propose to assign asymmetric weightings for ranking in the sense that incorrect placement of top-ranked items should yield higher penalty. Incorporation of the asymmetric weighting technique will further make our system to mimic human ranking strategy.
第一章 研究背景........................................1
第二章 簡介與相關研究...................................4
第三章 Pairwise排名學習................................7
3.1 排名評比(Ranking Measurement)....................8
3.1.1 Pairwise排名評比...............................8
3.1.2 分數式排名評比.................................10
3.2 排名學習........................................11
3.2.1 AdaBoost.....................................12
3.2.2 RankBoost....................................13
3.2.3 改良式RankBoost (RealRankBoost)...............16
3.2.4 排名同序的處理.................................18
第四章 排名學習器(Learners)............................21
4.1 排名函數(Ranking Function)......................21
4.1.1 Linear Weight Vector.........................22
4.1.2 Polynomial Weight Function...................24
4.1.3 Radial Basis Function........................28
4.2 排名方法（現有排名方法）..........................32
第五章 資料實測與效能分析...............................33
5.1 LETOR排名評測標準................................33
5.1.1 Dataset.......................................33
5.1.2 特徵值萃取(Feature extraction).................35
5.1.3 Data partitioning.............................38
5.1.4 效能評量.......................................39
5.2 實驗系統架構.....................................40
5.3 實驗結果與比較...................................42
第六章 非對稱加權排名...................................50
6.1 非對稱加權.......................................50
6.2 Pairwise上之非對稱加權方法........................52
6.3 加權函數(Weighting Function).....................54
6.4 實驗結果.........................................56
6.5 負向加權.........................................57
6.6 負向加權實驗......................................58
6.6.1 資料集(Dataset)................................58
6.6.2 實驗設計.......................................59
6.6.3 實驗結果.......................................60
第七章 結論與未來規劃....................................62
參考文獻................................................64
附錄....................................................66
[1]Thorsten Joachims, “Optimizing Search Engines using Clickthrough Data,” Proceedings of the ACM Conference on Knowledge Discovery and Data Mining (KDD), ACM, pp.133-142, 2002
[2]Zhe Cao, Tao Qin, Tie-Yan Liu, Ming-Feng Tsai, Hang Li, “Learning to Rank: From Pairwise Approach to Listwise Approach,” Proceedings of the 24th international conference on Machine learning (ICML), pp. 129-136, 2007
[3]Ping Li, Christopher J.C. Burges, Qiang Wu, “McRank: Learning to Rank Using Multiple Classification and Gradient Boosting,” Neural Information Processing Systems(NIPS), pp. 897-904, 2007
[4]Yoav Freund, Raj Iyer, Robert E. Schapire, Yoram Singer, “An Efficient Boosting Algorithm for Combining Preferences,” International Conference on Machine Learning, 1998
[5]Ming-Feng Tsai, Tie-Yan Liu, Tao Qin, Hsin-Hsi Chen, Wei-Ying Ma, “FRank: A Ranking Method with Fidelity Loss,” Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, pp. 383-390, 2007
[6]Jun Xu, Hang Li, “AdaRank: A Boosting Algorithm for Information Retrieval,” Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, pp. 391-398, 2007
[7]Christopher J.C. Burges, “A Tutorial on Support Vector Machines for Pattern Recognition,” Data Mining and Knowledge Discovery 2, pp. 121-167, 1998
[8]Robert E. Schapire, Yoram Singer, “Improved Boosting Algorithms Using Confidence-rated Predictions,” Machine Learning, 37(3), pp. 297-336, 1999
[9]Maurice George Kendall, “Rank Correlation Methods,” Hafner, 1955
[10]Frank Rosenblatt, “The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain,” Cornell Aeronautical Laboratory, Psychological Review, v65, No. 6, pp. 386-408, 1958
[11]Marvin Minsky, Seymour Papert, “Perceptrons” Neurocomputing: foundations of research, MIT Press, pp. 157-169, 1988
[12]Jooyoung Park, Irwin W Sandberg, “Universal Approximation Using Radial- Basis-Function Networks,” Neural Computation, MIT Press, pp. 246-257, 1991
[13]J.A. Leonard, M.A. Kramer, L.H. Ungar, “Using Radial Basis Functions to Approximate a Function and Its Error Bounds,” IEEE Transactions on Neural Networks, pp. 207-224, 1991
[14]Tie-Yan Liu, Tao Qin, Jun Xu, Wenying Xiong and Hang Li, “LETOR: Benchmark Dataset for Research on Learning to Rank for Information Retrieval,” LR4IR 2007, in conjunction with SIGIR 2007, 2007
[15]William Hersh, Chris Buckley, T. J. Leone, David Hickam, “OHSUMED: An Interactive Retrieval Evaluation and New Large Test Collection for Research,” Proceedings of the 17th Annual ACM SIGIR Conference, pp. 192-201, 1994
[16]Nick Craswell, David Hawking, “Overview of the TREC 2004 Web Track,” The 13th Text Retrieval Conference (TREC 2004), 2004
[17]Stephen E. Robertson, “Overview of the okapi projects,” Journal of Documentation, Vol. 53, No. 1, pp.3-7, 1997