A Bag Oversampling Approach for Class Imbalance in Multiple Instance Learning

  • Carlos MeraEmail author
  • Jose Arrieta
  • Mauricio Orozco-Alzate
  • John Branch
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9423)


Multiple Instance Learning (MIL) is a relatively new learning paradigm which allows to train a classifier with weakly labelled data. In spite that the community has been developing different methods to learn from this kind of data, there is little discussion on how to proceed when there is an imbalanced representation of the classes. The class imbalance problem in MIL is more complex compared with their counterpart in single-instance learning because it may occur at instance and/or bag level, or at both. Here, we propose an oversampling approach at bag level in order to improve the representation of the minority class. Experiments in nine benchmark data sets are conducted to evaluate the proposed approach.


Multiple instance learning Class imbalance Oversampling Bag oversampling 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Alba-Cabrera, E., Santana, R., Ochoa-Rodriguez, A., Lazo-Corts, M.: Finding typical testors by using an evolutionary strategy. In: Proceedings of the 5th Ibero American Symposium on Pattern Recognition, p. 267 (2000)Google Scholar
  2. 2.
    Alba-Cabrera, E., Ibarra-Fiallo, J., Godoy-Calderon, S.: A theoretical and practical framework for assessing the computational behavior of typical testor-finding algorithms. In: Ruiz-Shulcloper, J., Sanniti di Baja, G. (eds.) CIARP 2013, Part I. LNCS, vol. 8258, pp. 351–358. Springer, Heidelberg (2013) CrossRefGoogle Scholar
  3. 3.
    Diaz-Sanchez, G., Piza-Davila, I., Sanchez-Diaz, G., Mora-Gonzalez, M., Reyes-Cardenas, O., Cardenas-Tristan, A., Aguirre-Salado, C.: Typical testors generation based on an evolutionary algorithm. In: Yin, H., Wang, W., Rayward-Smith, V. (eds.) IDEAL 2011. LNCS, vol. 6936, pp. 58–65. Springer, Heidelberg (2011) CrossRefGoogle Scholar
  4. 4.
    Lazo-Cortés, M., Ruiz-Shulcloper, J., Alba-Cabrera, E.: An overview of the evolution of the concept of testor. Pattern Recognition 34(4), 753–762 (2001)CrossRefzbMATHGoogle Scholar
  5. 5.
    Lias-Rodríguez, A., Pons-Porrata, A.: BR: a new method for computing all typical testors. In: Bayro-Corrochano, E., Eklundh, J.-O. (eds.) CIARP 2009. LNCS, vol. 5856, pp. 433–440. Springer, Heidelberg (2009) CrossRefGoogle Scholar
  6. 6.
    Martnez-Trinidad, J.F., Guzmán-Arenas, A.: The logical combinatorial approach to pattern recognition, an overview through selected works. Pattern Recognition 34(4), 741–751 (2001)CrossRefGoogle Scholar
  7. 7.
    Sanchez-Diaz, G., Lazo-Cortes, M., Piza-Davila, I.: A fast implementation for the typical testor property identification based on an accumulative binary tuple. International Journal of Computational Intelligence Systems 5(6), 1025–1039 (2012)CrossRefGoogle Scholar
  8. 8.
    Sanchez-Diaz, G., Diaz-Sanchez, G., Mora-Gonzalez, M., Piza-Davila, I., Aguirre-Salado, C., Huerta-Cuellar, G., Reyes-Cardenas, O., Cardenas-Tristan, A.: An evolutionary algorithm with acceleration operator to generate a subset of typical testors. Pattern Recognition Letters 41, 34–42 (2014)CrossRefGoogle Scholar
  9. 9.
    Santiesteban-Alganza, Y., Pons-Porrata, A.: LEX: A new algorithm for calculating typical testors. Revista Ciencias Matematicas 21(1), 85–95 (2003)MathSciNetGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Carlos Mera
    • 1
    Email author
  • Jose Arrieta
    • 1
  • Mauricio Orozco-Alzate
    • 2
  • John Branch
    • 1
  1. 1.Universidad Nacional de ColombiaSede MedellínColombia
  2. 2.Universidad Nacional de ColombiaSede ManizalesColombia

Personalised recommendations