Advertisement

Multi-Classifier Adaptive Training: Specialising an Activity Recognition Classifier Using Semi-supervised Learning

  • Božidara Cvetković
  • Boštjan Kaluža
  • Mitja Luštrek
  • Matjaž Gams
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7683)

Abstract

When an activity recognition classifier is deployed to be used with a particular user, its performance can often be improved by adapting it to that user. To improve the classifier, we propose a novel semi-supervised Multi-Classifier Adaptive Training algorithm (MCAT) that uses four classifiers. First, the General classifier is trained on the labelled data available before deployment. Second, the Specific classifier is trained on a limited amount of labelled data specific to the new user in the current environment. Third, a domain-independent meta-classifier decides whether to classify a new instance with the General or Specific classifier. Fourth, another meta-classifier decides whether to include the new instance in the training set for the General classifier. The General classifier is periodically retrained, gradually adapting to the new user in the new environment where it is deployed. The results show that our new algorithm outperforms competing approaches and increases the accuracy of the initial activity recognition classifier by 12.66 percentage points on average.

Keywords

semi-supervised learning adaptation to the user MCAT activity recognition 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Cvetković, B., Kaluža, B., Luštrek, M., Gams, M.: Semi-supervised Learning for Adaptation of Human Activity Recognition Classifier to the User. In: Workshop on Space, Time and Ambient Intelligence, IJCAI 2011, Barcelona, pp. 24–29 (2011)Google Scholar
  2. 2.
    Zhu, X.: Semi-Supervised Learning Literature Survey. CS Technical Report. University of Wisconsin-Madison (2005)Google Scholar
  3. 3.
    Kuncheva, L.I., Whitaker, C.J., Duin, R.P.W.: Limits on the majority vote accuracy in classifier fusion. Pattern Analysis and Applications 8, 22–31 (2003)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Longstaff, B., Reddy, S., Estrin, D.: Improving activity classification for health applications on mobile devices using active and semi-supervised learning. In: 4th International Conference on Pervasive Computing Technologies for Healthcare, pp. 1–7. IEEE Press (2010)Google Scholar
  5. 5.
    Dasgupta, S.: Two faces of active learning. Theoretical Computer Science 412, 1767–1781 (2011)MathSciNetzbMATHCrossRefGoogle Scholar
  6. 6.
    Settles, B.: Active Learning Literature Survey. CS Technical Report. University of Wisconsin–Madison (2009)Google Scholar
  7. 7.
    Bicocchi, N., Mamei, M., Prati, A., Cucchiara, R., Zambonelli, F.: Pervasive Self-Learning with Multi-modal Distributed Sensors. In: 2nd IEEE International Conference on Self-Adaptive and Self-Organizing Systems, pp. 61–66. IEEE Press, Washington (2008)CrossRefGoogle Scholar
  8. 8.
    Frinken, V., Bunke, H.: Self-training Strategies for Handwriting Word Recognition. In: Perner, P. (ed.) ICDM 2009. LNCS, vol. 5633, pp. 291–300. Springer, Heidelberg (2009)CrossRefGoogle Scholar
  9. 9.
    Guzmán-Cabrera, R., Montes-y-Gómez, M., Rosso, P., Villaseñor-Pineda, L.: A Web-Based Self-training Approach for Authorship Attribution. In: Nordström, B., Ranta, A. (eds.) GoTAL 2008. LNCS (LNAI), vol. 5221, pp. 160–168. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  10. 10.
    Feng-Biao, G., Chun-Ting, Z.: ZCURVEV: a new self-training system for recognizing protein-coding genes in viral and phage genomes. BMC Bioinformatics 7, 9 (2006)CrossRefGoogle Scholar
  11. 11.
    Blum, A., Mitchell, T.: Combining labeled and unlabeled data with co-training. In: 11th Annual Conference on Computational Learning Theory, pp. 92–100. Morgan Kaufmann Press (1998)Google Scholar
  12. 12.
    Guan, D., Yuan, W., Lee, Y.K., Gavrilov, A., Lee, S.: Activity Recognition Based on Semi-supervised Learning. In: IEEE International Conference on Embedded and Real-Time Computing Systems and Applications, pp. 469–475. IEEE Press, Washington (2007)Google Scholar
  13. 13.
    Zhou, Y., Goldman, S.: Democratic Co-Learning. In:16th IEEE International Conference on Tools with Artificial Intelligence, pp. 594–602. IEEE Press, Washington (2004)CrossRefGoogle Scholar
  14. 14.
    Project Confidence, http://www.confidence-eu.org/
  15. 15.
    Dovgan, E., Luštrek, M., Pogorelc, B., Gradišek, A., Burger, H., Gams, M.: Intelligent elderly-care prototype for fall and disease detection from sensor data. Zdravniski Vestnik - Slovenian Medical Journal 80(11), 824–831 (2011)Google Scholar
  16. 16.
    Kaluža, B., Mirchevska, V., Dovgan, E., Luštrek, M., Gams, M.: An Agent-Based Approach to Care in Independent Living. In: de Ruyter, B., Wichert, R., Keyson, D.V., Markopoulos, P., Streitz, N., Divitini, M., Georgantas, N., Mana Gomez, A. (eds.) AmI 2010. LNCS, vol. 6439, pp. 177–186. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  17. 17.
  18. 18.
    Kaluža, B., Dovgan, E.: Glajenje trajektorij gibanja cloveškega telesa zajetih z radijsko tehnologijo. In: Proceedings of the 12th International Multiconference Information Society, IS 2009, Ljubljana, Slovenia, vol. A, pp. 97–100 (2009)Google Scholar
  19. 19.
    Luštrek, M., Kaluža, B.: Fall detection and activity recognition with machine learning. Informatica 33, 197–204 (2009)Google Scholar
  20. 20.
    Yanhua, C., Manjeet, R., Ming, D., Jing, H.: Non-negative matrix factorization for semi-supervised data clustering. Knowledge and Information Systems 17, 355–379 (2008)CrossRefGoogle Scholar
  21. 21.
    UC Irvine Machine Learning Repository, http://archive.ics.uci.edu/ml/
  22. 22.
    Platt, J.C.: Fast training of support vector machines using sequential minimal optimization. MIT Press, Cambridge (1999)Google Scholar
  23. 23.
    Ross, Q.J.: C4.5: Programs for Machine Learning. Morgan Kaufmann, San Francisco (1992)Google Scholar
  24. 24.
    Hall, M., Frank, E., Holems, G., Pfahringer, B., Reutemann, P., Witten, I.H.: The WEKA Data Mining Software: An Update. SIGKDD Explorations 11, 10–18 (2009)CrossRefGoogle Scholar
  25. 25.
    Breiman, L.: Random Forests. Machine Learning 45, 5–32 (2001)zbMATHCrossRefGoogle Scholar
  26. 26.
    Gjoreski, H.: Adaptive human activity/posture recognition and fall detection using wearable sensors. Master thesis, Jožef Stefan International Postgraduate School (2011)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Božidara Cvetković
    • 1
  • Boštjan Kaluža
    • 1
  • Mitja Luštrek
    • 1
  • Matjaž Gams
    • 1
  1. 1.Department of Intelligent SystemsJožef Stefan InstituteLjubljanaSlovenia

Personalised recommendations