Nothing Special   »   [go: up one dir, main page]
Skip to main content

Advertisement

Springer Nature Link
Log in

Self-labeled techniques for semi-supervised learning: taxonomy, software and empirical study

  • Survey Paper
  • Published:
Knowledge and Information Systems Aims and scope Submit manuscript

Abstract

Semi-supervised classification methods are suitable tools to tackle training sets with large amounts of unlabeled data and a small quantity of labeled data. This problem has been addressed by several approaches with different assumptions about the characteristics of the input data. Among them, self-labeled techniques follow an iterative procedure, aiming to obtain an enlarged labeled data set, in which they accept that their own predictions tend to be correct. In this paper, we provide a survey of self-labeled methods for semi-supervised classification. From a theoretical point of view, we propose a taxonomy based on the main characteristics presented in them. Empirically, we conduct an exhaustive study that involves a large number of data sets, with different ratios of labeled data, aiming to measure their performance in terms of transductive and inductive classification capabilities. The results are contrasted with nonparametric statistical tests. Note is then taken of which self-labeled models are the best-performing ones. Moreover, a semi-supervised learning module has been developed for the Knowledge Extraction based on Evolutionary Learning software, integrating analyzed methods and data sets.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

Similar content being viewed by others

Notes

  1. http://sci2s.ugr.es/keel/datasets.

  2. http://sci2s.ugr.es/SelfLabeled.

  3. http://www.keel.es.

References

  1. Zhu X, Goldberg AB (2009) Introduction to semi-supervised learning, 1st edn. Morgan and Claypool, San Rafael, CA

    MATH  Google Scholar 

  2. Witten IH, Frank E, Hall MA (2011) Data mining: practical machine learning tools and techniques, 3rd edn. Morgan Kaufmann, San Francisco

    Google Scholar 

  3. Zhu Y, Yu J, Jing L (2013) A novel semi-supervised learning framework with simultaneous text representing. Knowl Inf Syst 34(3):547–562

    Article  Google Scholar 

  4. Chapelle O, Schlkopf B, Zien A (2006) Semi-supervised learning, 1st edn. The MIT Press, Cambridge, MA

    Book  Google Scholar 

  5. Pedrycz W (1985) Algorithms of fuzzy clustering with partial supervision. Pattern Recognit Lett 3:13–20

    Article  Google Scholar 

  6. Zhao W, He Q, Ma H, Shi Z (2012) Effective semi-supervised document clustering via active learning with instance-level constraints. Knowl Inf Syst 30(3):569–587

    Article  Google Scholar 

  7. Chen K, Wang S (2011) Semi-supervised learning via regularized boosting working on multiple semi-supervised assumptions. IEEE Trans Pattern Anal Mach Intell 33(1):129–143

    Article  Google Scholar 

  8. Fujino A, Ueda N, Saito K (2008) Semisupervised learning for a hybrid generative/discriminative classifier based on the maximum entropy principle. IEEE Trans Pattern Anal Mach Intell 30(3):424–437

    Article  Google Scholar 

  9. Joachims T (1999) Transductive inference for text classification using support vector machines. In: Proceedings of 16th international conference on machine learning, Morgan Kaufmann, pp 200–209

  10. Blum A, Chawla S (2001) Learning from labeled and unlabeled data using graph mincuts. In: Proceedings of the eighteenth international conference on machine learning, pp 19–26

  11. Wang J, Jebara T, Chang S-F (2013) Semi-supervised learning using greedy max-cut. J Mac Learn Res 14(1):771–800

    MATH  MathSciNet  Google Scholar 

  12. Mallapragada PK, Jin R, Jain A, Liu Y (2009) Semiboost: boosting for semi-supervised learning. IEEE Trans Pattern Anal Mach Intell 31(11):2000–2014

    Article  Google Scholar 

  13. Yarowsky D (1995) Unsupervised word sense disambiguation rivaling supervised methods. In: Proceedings of the 33rd annual meeting of the association for computational linguistics, pp 189–196

  14. Li M, Zhou ZH (2005) SETRED: self-training with editing. In: Lecture notes in computer science (including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), vol 3518 LNAI, pp 611–621

  15. Blum A, Mitchell T (1998) Combining labeled and unlabeled data with co-training. In: Proceedings of the annual ACM conference on computational learning theory, pp 92–100

  16. Du J, Ling CX, Zhou ZH (2010) When does co-training work in real data? IEEE Trans Knowl Data Eng 23(5):788–799

    Article  Google Scholar 

  17. Sun S, Jin F (2011) Robust co-training. Int J Pattern Recognit Artif Intell 25(07):1113–1126

    Article  MathSciNet  Google Scholar 

  18. Jiang Z, Zhang S, Zeng J (2013) A hybrid generative/discriminative method for semi-supervised classification. Knowl-Based Syst 37:137–145

    Article  Google Scholar 

  19. Sun S (2013) A survey of multi-view machine learning. Neural Comput Appl 23(7–8):2031–2038

    Google Scholar 

  20. Zhou ZH, Li M (2005) Tri-training: exploiting unlabeled data using three classifiers. IEEE Trans Knowl Data Eng 17:1529–1541

    Article  Google Scholar 

  21. Li M, Zhou ZH (2007) Improve computer-aided diagnosis with machine learning techniques using undiagnosed samples. IEEE Trans Syst Man Cybern A Syst Hum 37(6):1088–1098

    Article  Google Scholar 

  22. Sun S, Shawe-Taylor J (2010) Sparse semi-supervised learning using conjugate functions. J Mach Learn Res 11:2423–2455

    MATH  MathSciNet  Google Scholar 

  23. Zhu X (2005) Semi-supervised learning literature survey. Technical report 1530, Computer Sciences, University of Wisconsin-Madison

  24. Chawla N, Karakoulas G (2005) Learning from labeled and unlabeled data: an empirical study across techniques and domains. J Artif Intell Res 23:331–366

    MATH  Google Scholar 

  25. Zhou Z-H, Li M (2010) Semi-supervised learning by disagreement. Knowl Inf Syst 24(3):415–439

    Article  Google Scholar 

  26. Alcalá-Fdez J, Sánchez L, García S, del Jesus MJ, Ventura S, Garrell JM, Otero J, Romero C, Bacardit J, Rivas VM, Fernández JC, Herrera F (2009) KEEL: a software tool to assess evolutionary algorithms for data mining problems. Soft Comput 13(3):307–318

    Article  Google Scholar 

  27. Demšar J (2006) Statistical comparisons of classifiers over multiple data sets. J Mach Learn Res 7:1–30

    MATH  MathSciNet  Google Scholar 

  28. García S, Fernández A, Luengo J, Herrera F (2010) Advanced nonparametric tests for multiple comparisons in the design of experiments in computational intelligence and data mining: experimental analysis of power. Inf Sci 180:2044–2064

    Article  Google Scholar 

  29. Triguero I, Sáez JA, Luengo J, García S, Herrera F (2013) On the characterization of noise filters for self-training semi-supervised in nearest neighbor classification, Neurocomputing (in press)

  30. Cover TM, Hart PE (1967) Nearest neighbor pattern classification. IEEE Trans Inf Theory 13(1):21–27

    Article  MATH  Google Scholar 

  31. Dasgupta S, Littman ML, McAllester DA (2001) Pac generalization bounds for co-training. In: Dietterich TG, Becker S, Ghahramani Z (eds) Advances in neural information processing systems. Neural information processing systems: natural and synthetic, vol 14. MIT Press, Cambridge, pp 375–382

  32. Quinlan JR (1993) C4.5 programs for machine learning. Morgan Kaufmann Publishers, San Francisco, CA

    Google Scholar 

  33. Efron B, Tibshirani RJ (1993) An Introduction to the bootstrap. Chapman & Hall, New York

    Book  MATH  Google Scholar 

  34. Goldman S, Zhou Y (2000) Enhancing supervised learning with unlabeled data. In: Proceedings of the 17th international conference on machine learning. Morgan Kaufmann, pp 327–334

  35. Alcalá-Fdez J, Fernandez A, Luengo J, Derrac J, García S, Sánchez L, Herrera F (2011) KEEL data-mining software tool: data set repository, integration of algorithms and experimental analysis framework. J Multiple-Valued Logic Soft Comput 17(2–3):255–277

    Google Scholar 

  36. Bennett K, Demiriz A, Maclin R (2002) Exploiting unlabeled data in ensemble methods. In: Proceedings of the ACM SIGKDD international conference on knowledge discovery and data mining, pp 289–296

  37. Zhou Y, Goldman S (2004) Democratic co-learning. In: IEEE international conference on tools with artificial intelligence, pp 594–602

  38. Deng C, Guo M (2006) Tri-training and data editing based semi-supervised clustering algorithm. In: Gelbukh A, Reyes-Garcia C (eds) MICAI 2006: advances in artificial intelligence, vol 4293 of lecture notes in computer science. Springer, Berlin, pp 641–651

  39. Wang J, Luo S, Zeng X (2008) A random subspace method for co-training. In: IEEE international joint conference on computational intelligence, pp 195–200

  40. Hady M, Schwenker F (2008) Co-training by committee: a new semi-supervised learning framework. In: IEEE international conference on data mining workshops, ICDMW ’08, pp 563–572

  41. Hady M, Schwenker F (2010) Combining committee-based semi-supervised learning and active learning. J Comput Sci Technol 25:681–698

    Article  Google Scholar 

  42. Hady M, Schwenker F, Palm G (2010) Semi-supervised learning for tree-structured ensembles of rbf networks with co-training. Neural Netw 23:497–509

    Article  Google Scholar 

  43. Yaslan Y, Cataltepe Z (2010) Co-training with relevant random subspaces. Neurocomputing 73(10–12):1652–1661

    Article  Google Scholar 

  44. Huang T, Yu Y, Guo G, Li K (2010) A classification algorithm based on local cluster centers with a few labeled training examples. Knowl-Based Syst 23(6):563–571

    Article  Google Scholar 

  45. Halder A, Ghosh S, Ghosh A (2010) Ant based semi-supervised classification. In: Proceedings of the 7th international conference on swarm intelligence, ANTS’10, Springer, Berlin, Heidelberg, pp 376–383

  46. Wang Y, Xu X, Zhao H, Hua Z (2010) Semi-supervised learning based on nearest neighbor rule and cut edges. Knowl-Based Syst 23(6):547–554

    Article  Google Scholar 

  47. Deng C, Guo M (2011) A new co-training-style random forest for computer aided diagnosis. J Intell Inf Syst 36:253–281. doi:10.1007/s10844-009-0105-8

    Article  Google Scholar 

  48. Nigam K, Mccallum A, Thrun S, Mitchell T (2000) Text classification from labeled and unlabeled documents using em. Mach Learn 39(2):103–134

    Article  MATH  Google Scholar 

  49. Tang X-L, Han M (2010) Semi-supervised Bayesian artmap. Appl Intell 33(3):302–317

    Article  MathSciNet  Google Scholar 

  50. Joachims T (2003) Transductive learning via spectral graph partitioning. In: Proceedings of twentieth international conference on machine learning, vol 1, pp 290–297

  51. Belkin M, Niyogi P, Sindhwani V (2006) Manifold regularization: a geometric framework for learning from labeled and unlabeled examples. J Mach Learn Res 7:2399–2434

    MATH  MathSciNet  Google Scholar 

  52. Xie B, Wang M, Tao D (2011) Toward the optimization of normalized graph Laplacian. IEEE Trans Neural Netw 22(4):660–666

    Article  Google Scholar 

  53. Burges C (1998) A tutorial on support vector machines for pattern recognition. Data Min Knowl Discov 2(2):121–167

    Article  Google Scholar 

  54. Chapelle O, Sindhwani V, Keerthi SS (2008) Optimization techniques for semi-supervised support vector machines. J Mach Learn Re. 9:203–233

    MATH  Google Scholar 

  55. Adankon M, Cheriet M (2010) Genetic algorithm-based training for semi-supervised svm. Neural Comput Appl 19:1197–1206

    Article  Google Scholar 

  56. Tian X, Gasso G, Canu S (2012) A multiple kernel framework for inductive semi-supervised svm learning. Neurocomputing 90:46–58

    Article  Google Scholar 

  57. Sugato B, Raymond JM (2003) Comparing and unifying search-based and similarity-based approaches to semi-supervised clustering. In: Proceedings of the ICML-2003 workshop on the continuum from labeled to unlabeled data in machine learning and data mining, pp 42–49

  58. Yin X, Chen S, Hu E, Zhang D (2010) Semi-supervised clustering with metric learning: an adaptive kernel method. Pattern Recognit 43(4):1320–1333

    Article  MATH  Google Scholar 

  59. Grira N, Crucianu M, Boujemaa N (2004) Unsupervised and semi-supervised clustering: a brief survey. In: A review of machine learning techniques for processing multimedia content. Report of the MUSCLE European network of excellence FP6

  60. Freund Y, Seung HS, Shamir E, Tishby N (1997) Selective sampling using the query by committee algorithm. Mach Learn 28:133–168

    Article  MATH  Google Scholar 

  61. Muslea I, Minton S, Knoblock C (2002) Active + semi-supervised learning = robust multi-view learning. In: Proceedings of ICML-02, 19th international conference on machine learning, pp 435–442

  62. Zhang Q, Sun S (2010) Multiple-view multiple-learner active learning. Pattern Recognit 43(9):3113–3119

    Google Scholar 

  63. Yu H (2011) Selective sampling techniques for feedback-based data retrieval. Data Min Knowl Discov 22(1–2):1–30

    Article  MATH  MathSciNet  Google Scholar 

  64. Belhumeur P, Hespanha J, Kriegman D (1997) Eigenfaces vs. fisherfaces: recognition using class specific linear projection. IEEE Trans Pattern Anal Mach Intell 19(7):711–720

    Article  Google Scholar 

  65. Hastie T, Tibshirani R, Friedman J (2001) The elements of statistical learning: data mining, inference and prediction, 2nd edn. Springer, Berlin

    Book  Google Scholar 

  66. Song Y, Nie F, Zhang C, Xiang S (2008) A unified framework for semi-supervised dimensionality reduction. Pattern Recognit 41(9):2789–2799

    Article  MATH  Google Scholar 

  67. Li Y, Guan C (2008) Joint feature re-extraction and classification using an iterative semi-supervised support vector machine algorithm. Mach Learn 71:33–53

    Article  Google Scholar 

  68. Liu H, Motoda H (eds) (2007) Computational methods of feature selection. Chapman &Hall/CRC data mining and knowledge discovery series. Chapman & Hall/CRC, Boca Raton, FL

  69. Zhao J, Lu K, He X (2008) Locality sensitive semi-supervised feature selection. Neurocomputing 71(10–12):1842–1849

    Article  Google Scholar 

  70. Gregory PA, Gail AC (2010) Self-supervised ARTMAP. Neural Netw 23:265–282

    Article  Google Scholar 

  71. Cour T, Sapp B, Taskar B (2011) Learning from partial labels. J Mach Learn Res 12:1501–1536

    MATH  MathSciNet  Google Scholar 

  72. Joshi A, Papanikolopoulos N (2008) Learning to detect moving shadows in dynamic environments. IEEE Trans Pattern Anal Mach Intell 30(11):2055–2063

    Article  Google Scholar 

  73. Ben-David A (2007) A lot of randomness is hiding in accuracy. Eng Appl Artif Intell 20:875–885

    Article  Google Scholar 

  74. Alpaydin E (2010) Introduction to machine learning, 2nd edn. MIT Press, Cambridge, MA

    MATH  Google Scholar 

  75. Asuncion A, Newman D (2007) UCI machine learning repository. http://www.ics.uci.edu/mlearn/MLRepository.html

  76. Wu X, Kumar V (eds) (2009) The top ten algorithms in data mining. Chapman & Hall/CRC data mining and knowledge discovery. Chapman & Hall/CRC, Boca Raton, FL

  77. Aha DW, Kibler D, Albert MK (1991) Instance-based learning algorithms. Mach Learn 6(1):37–66

    Google Scholar 

  78. John GH, Langley P (2001) Estimating continuous distributions in Bayesian classifiers. In: Proceedings of the eleventh conference on uncertainty in artificial intelligence. Morgan Kaufmann, San Mateo, pp 338–345

  79. Vapnik VN (1998) Statistical learning theory. Wiley-Interscience, London

    MATH  Google Scholar 

  80. Platt JC (1999) Fast training of support vector machines using sequential minimal optimization. MIT Press, Cambridge, MA

    Google Scholar 

  81. García S, Herrera F (2008) An extension on statistical comparisons of classifiers over multiple data sets for all pairwise comparisons. J Mach Learn Res 9:2677–2694

    MATH  Google Scholar 

  82. Sheskin DJ (2011) Handbook of parametric and nonparametric statistical procedures, 5th edn. Chapman & Hall/CRC, Boca Raton, FL

    MATH  Google Scholar 

  83. Friedman M (1937) The use of ranks to avoid the assumption of normality implicit in the analysis of variance. J Am Stat Assoc 32:675–701

    Article  Google Scholar 

  84. Bergmann G, Hommel G (1988) Improvements of general multiple test procedures for redundant systems of hypotheses. In: Bauer P, Hommel G, Sonnemann E (eds) Multiple hypotheses testing. Springer, Berlin pp 100–115

  85. Yang Y, Webb G (2009) Discretization for naive-Bayes learning: managing discretization bias and variance. Mac Learn 74(1):39–74

    Article  Google Scholar 

  86. García S, Luengo J, Saez JA, López V, Herrera F (2013) A survey of discretization techniques: taxonomy and empirical analysis in supervised learning. IEEE Trans Knowl Data Eng 25(4):734–750

    Article  Google Scholar 

  87. Jolliffe IT (1986) Principal component analysis. Springer, Berlin

    Book  Google Scholar 

Download references

Acknowledgments

This work is supported by the Research Projects TIN2011-28488, TIC-6858 and P11-TIC-7765.

Author information

Authors and Affiliations

  1. Department of Computer Science and Artificial Intelligence, Research Center on Information and Communications Technology (CITIC-UGR), University of Granada, 18071 , Granada, Spain

    Isaac Triguero & Francisco Herrera

  2. Department of Computer Science, University of Jaén, 23071 , Jaén, Spain

    Salvador García

Authors
  1. Isaac Triguero
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Salvador García
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Francisco Herrera
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Isaac Triguero.

Appendix

Appendix

As a consequence of this work, we have developed a complete SSL framework which has been integrated into the Knowledge Extraction based on Evolutionary Learning (KEEL) toolFootnote 3 [26]. This research tool is an open-source software, written in Java, that supports data management and the design of experiments. Until now, KEEL has paid special attention to the implementation of supervised and unsupervised learning, clustering, pattern mining and so on. Nevertheless, it did not offer support for SSL. We integrated a new SSL module into this software.

The main characteristics of this module are as follows:

  • All the data sets involved in the experimental study have been included into this module and can be used for new experiments. These data sets are composed of three files for each partition: training, transductive and test partitions. The former is composed of labeled and unlabeled instances (labeled as “unlabeled”). Transductive partition contains the real class of unlabeled instances and the latter collect the test instances. These data sets are included in the KEEL-data set repository and are static, ensuring that further experiments carried out will no longer be dependent on particular data partitions.

  • It allows the design of SSL experiments which generate all the XML scripts and a JAR program for running it, by creating a zip file for an off-line run. The SSL module is designed for experiments containing multiple data sets and algorithms connected among themselves to obtain the desired experimental setup. The parameters configuration of the methods is also customizable as well as the number of executions, validation scheme and so on. Figure 13 shows a snapshot of an experiment with three analyzed self-labeled methods and the customization of the parameters of the algorithm APSSC. Note that every method could be executed apart from the KEEL tool with an appropriate configuration file.

  • Special care has been taken to allow a researcher to be able to use this module to assess the relative effectiveness of his own procedures. Guidelines about how to integrate a method into KEEL can be found in [35].

The KEEL version with the SSL module is available on the associated Web site.

Fig. 13
figure 13

A snapshot of the semi-supervised learning module for KEEL

Full size image

Rights and permissions

Reprints and permissions

About this article

Cite this article

Triguero, I., García, S. & Herrera, F. Self-labeled techniques for semi-supervised learning: taxonomy, software and empirical study. Knowl Inf Syst 42, 245–284 (2015). https://doi.org/10.1007/s10115-013-0706-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10115-013-0706-y

Keywords

Search

Navigation