计算机科学 ›› 2017, Vol. 44 ›› Issue (12): 255-259.doi: 10.11896/j.issn.1002-137X.2017.12.046
杨潇,崔超然,王帅强
YANG Xiao, CUI Chao-ran and WANG Shuai-qiang
摘要: 在排序学习中引入特征选择可以提高学习的效率和准确率。出于对选择速度的考虑,当前的研究主要从特征选择的角度出发,根据特征对排序的作用和特征之间的相似性选择对排序区分度最大的特征集合。由于特征大都是人工归纳的,因此特征和特征之间难免存在重叠和冗余。为了减少特征之间的冗余,从特征生成的角度出发,对现有特征进行矩阵分解,从而生成新的特征集。考虑到使用奇异值分解(Singular Value Decomposition SVD)等方法进行矩阵分解时不能综合考虑排序结果对特征的影响,基于特征矩阵对排序的效果、特征矩阵与原矩阵之间的差距来构造优化算法,提出了一种基于矩阵分解的排序学习优化方法,并根据该优化方法设计了排序学习特征选择算法MFRank。实验中使用映射随机梯度下降法近似求得优化问题的最优值,在公开测试集MQ2008上的结果显示,所提MFRank方法获得了与当前最优的特征选择方法即RankBoost和RankSVM-Struct等排序算法相当的结果。
[1] HUANG Z H,ZHANG J W,TIAN C Q,et al.Survey on Lear-ning-to-Rank Based Recommendation Algorithms [J].Journal of Software,2016,27(3):691-713.(in Chinese) 黄震华,张佳雯,田春岐,等.基于排序学习的推荐算法研究综述[J].软件学报,2016,7(3):691-713. [2] GUO L,MA J,CHEN Z,et al.Learning to Recommend with Social Contextual Information from Implicit Feedback[J].Soft Computing,2015,19(5):1351-1362. [3] LIU L,LU X,LIAO Y,et al.Improving Retrieval of PlaneGeometry Figure with Learning to Rank[J].Pattern Recognition Letters,2016,83(3):423-429. [4] http://research.microsoft.com/en-us/projects/ mslr/ feature.aspx. [5] PAN F,CONVERSE T,AHN D,et al.Greedy and Randomized Feature Selection for Web Search Ranking[C]∥Proceedings of 11th IEEE International Conference on Computer and Information Technology.Sydney:IEEE Press,2011:436-442. [6] JRVELIN K,KEKLINEN J.Cumulated Gain-based Eva-luation of IR Techniques[J].ACM Transactions on Information Systems,2002,20(4):422-446. [7] YATES R B,NETO B R.Modern Information Retrieval[M].New York:Addison Weskey,1999. [8] GENG X,LIU T,QIN T.Feature Selection for Ranking[C]∥Sigir:International ACM SIGIR Conference on Research & Development in Information Retrieval.New York:ACM,2007:407-414. [9] DANG V,CROFT W B.Feature Selection for Document Ran-king using Best First Search and Coordinate Ascent [C]∥Proceedings of SIGIR Workshop Feature Generation and Selection of Information Retrieval.2010:1-5. [10] LI C,SHAO L,XU C S,et al.Feature selection under learning to rank model for multimedia retrieval[C]∥Proceedings of 2nd International Conference of Internet Multimedia Computer Ser-vice.2010:69-72. [11] DA SILVA S R F,RIBEIRO M X,NETO J E S B,et al.Improving the Ranking Quality of Medical Image Retrieval using a Genetic Feature Selection Method[J].Decision Support System,2011,5(4):810-820. [12] LAI H,PAN Y,TANG Y,et al.FSMRank:Feature Selection Algorithm for Learning to Rank[J].IEEE Transaction on Neural Networks and Learning Systems,2013,24(6):940-952. [13] HUA G C,ZHANG M,KUANG D,et al.Feature AnalysisMethods for Learning to Rank[J].Computer Engineering and Applications,2011,47(17):122-127.(in Chinese) 花贵春,张敏,邝达,等.面向排序学习的特征分析的研究[J].计算机工程与应用,2011,47(17):122-127. [14] LIN Y.Research of Learning to Rank in Information Retrieval[D].Dalian:Dalian University of Technology,2012.(in Chinese) 林原.信息检索中排序学习方法的研究[D].大连:大连理工大学,2012. [15] WU C G,LIANG Y Y,SUN Y F,et al.On the Equivalence of SVD and PCA[J].Chinese Journal of Computers,2004,7(2):286-288.(in Chinese) 吴春国,梁艳养,孙延风,等.关于SVD与PCA等价性的研究[J].计算机学报,2004,7(2):286-288. [16] ZHANG C.Research on Matrix Factorization Based Collaborative Filtering Recommendation Algorithms[D].Changchun:Jinlin University,2013.(in Chinese) 张川.基于矩阵分解的协同过滤推荐算法研究[D].长春:吉林大学,2013. [17] HILLERMEIER C.Nonlinear Multiobjective Optimization[M].Birkhaüser Verlag,Kluwer Academic Publishers,2001. [18] RENDLE S,FREUDENTHALER C,GANTNER Z,et al.Bayesian Personalized Ranking from Implicit Feedback[C]∥Proceedings of the Twenty-fifth Conference on Uncertainty in Artificial Intelligence.Arlington:AUAI Press,2009:452-461. [19] QIN T,LIU T,XU J,et al.LETOR:A Benchmark Collectionfor Research on Learning to Rank for Information Retrieval[J].Information Retrieval Journal,2010,13(4):346-374. [20] CAO Z,QIN T,LIU T,et al.Learning to Rank:From Pairwise Approach to Listwise Approach[C]∥Proceedings of 24th International Conference of Machine Learning.New York:ACM,2007:129-136. [21] JOACHIMS T.Training Linear SVMs in Linear Time[C]∥Proceedings of 12th ACM SIGKDD International Conference of Knowledge Discovery Data Mining.New York:ACM,2006:217-226. [22] FREUND Y,IYER R,SCHAPIRE R,et al.An Efficient Boosting Algorithm for Combining Preferences[J].Journal of Machine Learning Research,2003,4(11):933-969. [23] XU J,LI H.AdaRank:A Boosting Algorithm for Information Retrieval[C]∥Proceedings of 30th Annual International ACM SIGIR Conference Research & Develop in Information Retrie-val.New York,ACM,2007:391-398. |
No related articles found! |
|