Multi-Label Classification

From KDIS Research Group - Wiki

Jump to: navigation, search

Contents


[edit] Multi-Label Classification (MLC)

Multi-label classification is a classification paradigm that allows solving problems where patterns can be associated to more than one label. These are problems of increasing actuality, like document classification, and can be tackled with a more efficient approach than classical classification, which only allows one label per pattern.

Most of classification problems associate one label per pattern, li, from a set of disjoint labels, L. If such label set has only two possible values, representing if the pattern is a positive example of the class, it is called binary classification (|L|=2) while in other case it is called multi-class classification (|L|>2).

Nevertheless, this is not the only possible scenario, because there are a lot of current problems where a pattern can have associated not just one, but a set of class labels, Y Í L. In such cases it is called multi-label problem. Among the main multi-label problems text categorization, semantic scene and sounds classification and functional genes and protein classification should be stressed.

There are basically two main approaches to deal with multi-label problems: transformation problem methods and algorithm adaptation methods. The former consists basically in transforming a multi-label dataset into several binary classification problems and then applying a classical (single-label) algorithm. The latter tries to extend a specific classical algorithm to work directly with multi-label information

Our work is focused in the application of bioinspired algorithms to multi-label classification. We can highlight our GEP-MLC (Gene Expression Programming – Multi Label Classification) algorithm which extends the definition of Ferreira’s Gene Expression Programming in order to be able to work with multi-label data.

[edit] Resources

[edit] Members

  • Sebastián Ventura Soto
  • Eva Lucrecia Gibaja Galindo
  • José Luis Ávila Jiménez
  • Marco Antonio Barrón Ortiz
  • Oscar G. Reyes Pupo
  • Jose María Moyano Murillo

[edit] Our main contributions

  • Journal Papers
    • Jose M. Moyano, Eva L. Gibaja, and Sebastián Ventura. MLDA: A tool for analyzing multi-label datasets. Knowledge-Based Systems, 121:1 -- 3, 2017.
    • Alberto Cano, José María Luna, Eva L. Gibaja, and Sebastián Ventura. Laim discretization for multi-label data. Information Sciences, 330:370 -- 384, 2016.
    • Eva L. Gibaja, Jose M. Moyano, and Sebastián Ventura. An ensemble-based approach for multi-view multi-label classification. Progress in Artificial Intelligence, 5(4):251--259, 2016.
    • Oscar Reyes, Carlos Morell, and Sebastián Ventura. Effective lazy learning algorithm based on a data gravitation model for multi-label learning. Information Sciences, 340:159 -- 174, 2016.
    • Eva Gibaja and Sebastián Ventura. A tutorial on multilabel learning. ACM Comput. Surv., 47(3):52:1--52:38, 2015.
    • Oscar Reyes, Carlos Morell, and Sebastián Ventura. Scalable extensions of the ReliefF algorithm for weighting and selecting features on the multi-label learning context. Neurocomputing, 161:168 -- 182, 2015.
    • Eva Gibaja and Sebastián Ventura. Multi-label learning: a review of the state of the art and ongoing research. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 4(6):411--444, 2014.
    • Oscar Gabriel Reyes Pupo, Carlos Morell, and Sebastián Ventura. Evolutionary feature weighting to improve the performance of multi-label lazy algorithms. Integrated Computer-Aided Engineering, 21(4):339--354, 2014.
    • Alberto Cano, Amelia Zafra, Eva L. Gibaja, and Sebastián Ventura. A Grammar-Guided Genetic Programming Algorithm for Multi-Label Classification, pages 217--228. Springer Berlin Heidelberg, Berlin, Heidelberg, 2013.
    • O. G. Reyes, C. Morell & S. Ventura. Learning similarity metric to improve the performance of the multi-label lazy ranking algorithms. Integrated Computer-Aided Engineering, (submitted 2013).
    • J. L. Ávila, E. Gibaja , A. Zafra, S. Ventura. A Gene Expression Programming Algorithm for Multi-Label Classification. Journal of Multiple-Valued Logic and Soft Computing, 17(2-3), 183-206, 2011.
  • Conferences
    • Jose M Moyano, Eva L Gibaja, and Sebastián Ventura. Una herramienta para analizar conjuntos de datos multi-etiqueta. In V Congreso Español de Informática (CEDI 2016). VIII Simposio Teoría y Aplicaciones de Minería de Datos (TAMIDA 2016), pages 857--866, 2016.
    • Eva Gibaja, Jose M Moyano, and Sebastian Ventura. Combinación de vistas para clasificación multi-etiqueta: estudio preliminar. In XVI Conferencia de la Asociación Española para la Inteligencia Artificial, pages 759--768, 2015.
    • Jose M Moyano, Eva L Gibaja, and Sebastián Ventura. Diseño automático de multi-clasificadores basados en proyecciones de etiquetas. In XVI Conferencia de la Asociación Española para la Inteligencia Artificial (CAEPIA 2015), pages 355--365, 2015.
    • Jose M Moyano, Eva L Gibaja, Alberto Cano, Jose M Luna, and Sebastián Ventura. Algoritmo evolutivo para optimizar ensembles de clasificadores multi-**Oscar Gabriel Reyes Pupo, Carlos Morell, and Sebastián Ventura Soto. ReliefF-ML: An Extension of ReliefF Algorithm to Multi-label Learning. In CIARP 2013, pages 528--535. Springer Berlin Heidelberg, 2013.
    • O. G. Reyes, C. Morell, S. Ventura. Learning Similarity Metric to improve the performance of Lazy Multi-label Ranking Algorithms. Twelve International Conference on Intelligent System Design and Applications (ISDA 2012), Cochi, India, November 2012.
    • José Luis Ávila-Jiménez, Eva Gibaja, and Sebastián Ventura. Evolving Multi-label Classification Rules with Gene Expression Programming: A Preliminary Study. In HAIS 1010, pages 9--16, 2010.
    • E. Gibaja, M. Victoriano, J.L. Avila and S. Ventura. A TDIDT Technique For Multi-Label Classification. Tenth International Conference on Intelligent System Design and Applications (ISDA 2010), El Cairo, Egypt, November 2010.
    • Avila J.L., Gibaja, E. L., Ventura, S.: Multi-label classification with gene expression programming. Proceedings of the fourth International Conference on Hybrid Artificial Intelligent Systems (HAIS09), LNAI 5572, 629-637, ISBN: 978-3-642-02318-7.
    • Avila J.L., Gibaja, E. L., Zafra, A., Ventura, S.: A niching algorithm to learn discriminant functions with multi-label classification. Proceedings of the tenth International Conference on Intelligent Data Engineering and Automated Learning (IDEAL09), LNCS 5588, 570-577, ISBN: 978-3-642-04393-2

[edit] Other basic references

  • Chan A., Freitas, A.: A new ant colony algorithm for multi-label classification with applications in bioinfomatics. In GECCO ’06: Proceedings of the 8th annual conference on Genetic and evolutionary computation, (2006) 27–34, New York, NY, USA. ACM Press.
  • Clare, A., King, R.D.: Knowledge discovery in multi-label phenotype data. LNCS 2168 (2001)
  • Crammer, K., Singer, Y.: A family of additive online algorithms for category ranking. Journal of Machine Learning Research 3 (2003) 1025{1058
  • De Comité, F., Gilleron, R., Tommasi, M: Learning multi-label alternating decision trees from texts and data (2003) 251–274
  • Elisseeff, A., Weston, J.: A kernel method for multi-labelled classification. Advances in Neural Information Processing Systems, (2001) 14:681–687.
  • Fürnkranz, J., Hüllermeier, E., Loza, E, Brinker, K: Multilabel classification via calibrated label ranking. Machine Learning, (2008) 73(2):133–153

[edit] Multi-Label Classification Bibliography

Personal tools