Published
Using hybrid associative classifier with translation (HACT) for studying imbalanced data sets
Estudio de conjuntos de datos desbalanceados usando un modelo asociativo con traslación de ejes
DOI:
https://doi.org/10.15446/ing.investig.v32n1.28522Keywords:
data set, associative model, under sampling, class imbalance, pre-processing (en)Modelo asociativo, bajo-muestreo, clase des-balanceada, pre-procesamiento (es)
Class imbalance may reduce the classifier performance in several recognition pattern problems. Such negative effect is more notable with least represented class (minority class) Patterns. A strategy for handling this problem consisted of treating the classes included in this problem separately (majority and minority classes) to balance the data sets (DS). This paper has studied high sensitivity to class imbalance shown by an associative model of classification: hybrid associative classifier with translation (HACT); imbalanced DS impact on associative model performance was studied. The convenience of using sub-sampling methods for decreasing imbalanced negative effects on associative memories was analysed. This proposal's feasibility was based on experimental results obtained from eleven real-world datasets.
En diversos problemas de reconocimiento de patrones, se ha observado que el desequilibrio de clases puede disminuir el desempeño del clasificador, principalmente en los patrones de las clases minoritarias. Una estrategia para resolver el problema del desbalance, consiste en tratar por separado las clases incluidas en el problema (clase minoritaria o mayoritaria), a fin de equilibrar los conjuntos de datos. En este sentido, la motivación del presente artículo estriba en el hecho de que el modelo asociativo visto como Clasificador Híbrido Asociativo con Traslación (CHAT), es muy sensible al desbalance de las clases. Por ello, se analiza el impacto que los conjuntos de datos desbalanceados pueden tener sobre el rendimiento del CHAT. Adicionalmente, se analiza la conveniencia de utilizar métodos de bajo-muestreo para disminuir los efectos negativos que el modelo asociativo pueda sufrir. La viabilidad de este estudio se sustenta con los resultados experimentales obtenidos de once conjuntos de datos reales. Finalmente, el presente trabajo se considera como una investigación analítica-sintética.
References
Aldape-Pérez, M., Implementación de los modelos ALFA-BETA con lógica reconfigurable., MSc Computer Engineering thesis (digital systems), Centro de Investigación en Computación, IPN, 2007. pp. 6-16.
Álvarez, M., Estadística., ISBN 84-7485-327-3, Universidad de Deusto, Bilbao, 1994, pp.51-63.
Barandela, R., Cortés, N., Palacios, A., The nearest neighbour rule and the reduction of the training sample size., In Proceedings of the 9th Spanish Symposium on Pattern Recognition and Image Analysis, Universitat Jaume I, Benicasim, Spain, 2001, pp. 103-108.
Barandela, R., Hernández, J.K., Sánchez, J.S., Ferri, F.J., Imbalanced training set reduction and feature selection through genetic optimization., Proceeding of the 2005 conference on Artificial Intelligence Research and Development, ACM DL, Amsterdam, The Netherlands, 2005, pp. 215-222.
Batista, G. E. A. P.A., Carvalho, A. C. P. L. F., Monard, M. C., Applying one-sided selection to unbalanced datasets., Lecture Notes in Artificial Intelligence, Vol. 1793, 2000, pp. 315-325.
Cohen, G., Hilario, M., Hugonnet, S., Geissbuhler, A., Learning from Imbalanced data in surveillance of nosocomial infection. Artificial Intelligence in Medicine, ElSEVIER, Vol. 37, 2006, pp. 7-18.
Chawla, V. N., Bowyer, K. W., Hall, L. O., Kegelmeyer, W. P., SMOTE: Synthetic minority over-sampling technique., Journal of Artificial Intelligence Research, Vol. 16, 2002, pp. 321-357.
Daskalaki, S., Kopanas, I., Avouris, N., Evaluation of classifiers for an uneven class distribution problem., Applied Artificial Intelligence, Vol. 20, 2006, pp. 1-37.
Ezawa, K. J., Singh, M., Norton, S. W., Learning goal oriented Bayesian networks for telecommunications risk management., Machine Learning, Proceedings of the 13th International Conference, Ed. Morgan Kaufmann, Bari, Italy,1996, pp. 139-147.
Hart, P. E., The condensed nearest neighbour rule., IEEE Transactions on Information Theory, Vol. 14, 1968, pp. 515-516.
Huang, Y. M., Hung, C. M., Jiau, H. C., Evaluation of neural networks and data mining methods on a credit assessment task for class imbalance problem, Nonlinear Analysis: Real World Applications., Vol. 7, 2006, pp. 720-757.
Japkowicz, N., Stephen, S., The class imbalance problem: A systematic study, Intelligent Data Analysis., Vol. 6, 2002, pp. 429-449.
Kuncheva, L. O., Jain, L. C., Nearest neighbour classifier: simultaneous editing and feature selection., Pattern Recognition Letters, Vol. 20, 1999, pp. 1149-1156.
Prati, R. C., Batista, G. E. A. P. A., Monard, M. C., Class imbalance versus class overlapping: An analysis of a learning system behaviour., Lecture Notes in Computer Science, Vol. 2972, 2004a, pp. 312-321.
Prati, R. C., Batista, G. E. A. P. A., Monard, M. C., Learning with class skews and small disjoints. Proceedings of the 17th Brazilian Symposium on Artificial Intelligence, Ed. Springer, São Luís, Maranhão - Brazil, 2004b, pp. 1119-1139.
Ranawana, R., Palade, V., A new measure for classifier performance evaluation., Proceedings of IEEE Congress on Evolutionary Computation, IEEE, Vancouver, BC, 2006, pp. 2254-2261.
Santiago, R., Clasificador híbrido de patrones basado en la Lern-matrix de Steinbuch y el linear associator de Anderson Kohonen., MSc Computer Science thesis Centro de Investigación en Computación, IPN, 2003.
Tan, S., Neighbour-weighted Knearest neighbour for unbalanced text corpus, Expert Systems Applications., Vol. 28, 2005, pp. 667-671.
Weiss, G. M., Mining with rarity: a unifying framework., ACM SIGKDD Explorations Newsletter, Vol. 6, 2004, pp. 7-19.
Wilson, L., Asymptotic properties of nearest neighbour rules using edited data., IEEE Transactions on Systems, Man and Cybernetics, Vol. 2,1972, pp. 408-421.
Dimensions
PlumX
Article abstract page views
Downloads
How to Cite
License
Copyright (c) 2012 Laura Cleofas Sánchez, Magali Guzmán Escobedo, Rosa María Valdovinos Rosas, Cornelio Yáñez Márquez, Oscar Camacho Nieto
This work is licensed under a Creative Commons Attribution 4.0 International License.
The authors or holders of the copyright for each article hereby confer exclusive, limited and free authorization on the Universidad Nacional de Colombia's journal Ingeniería e Investigación concerning the aforementioned article which, once it has been evaluated and approved, will be submitted for publication, in line with the following items:
1. The version which has been corrected according to the evaluators' suggestions will be remitted and it will be made clear whether the aforementioned article is an unedited document regarding which the rights to be authorized are held and total responsibility will be assumed by the authors for the content of the work being submitted to Ingeniería e Investigación, the Universidad Nacional de Colombia and third-parties;
2. The authorization conferred on the journal will come into force from the date on which it is included in the respective volume and issue of Ingeniería e Investigación in the Open Journal Systems and on the journal's main page (https://revistas.unal.edu.co/index.php/ingeinv), as well as in different databases and indices in which the publication is indexed;
3. The authors authorize the Universidad Nacional de Colombia's journal Ingeniería e Investigación to publish the document in whatever required format (printed, digital, electronic or whatsoever known or yet to be discovered form) and authorize Ingeniería e Investigación to include the work in any indices and/or search engines deemed necessary for promoting its diffusion;
4. The authors accept that such authorization is given free of charge and they, therefore, waive any right to receive remuneration from the publication, distribution, public communication and any use whatsoever referred to in the terms of this authorization.
