Article Contents
Article Contents

# Selective further learning of hybrid ensemble for class imbalanced increment learning

• Incremental learning has been investigated by many researchers. However, only few works have considered the situation where class imbalance occurs. In this paper, class imbalanced incremental learning was investigated and an ensemble-based method, named Selective Further Learning (SFL) was proposed. In SFL, a hybrid ensemble of Naive Bayes (NB) and Multilayer Perceptrons (MLPs) were employed. For the ensemble of MLPs, parts of the MLPs were selected to learning from the new data set. Negative Correlation Learning (NCL) with Dynamic Sampling (DyS) for handling class imbalance was used as the basic training method. Besides, as an additive model, Naive Bayes was employed as an individual of the ensemble to learn the data sets incrementally. A group of weights (with the number of the classes as the length) are updated for every individual of the ensemble to indicate the 'confidence' of the individual learning about the classes. The ensemble combines all of the individuals by weighted average according to the weights. Experiments on 3 synthetic data sets and 10 real world data sets showed that SFL was able to handle class imbalance incremental learning and outperform a recently related approach.

Mathematics Subject Classification: Primary: 58F15, 58F17; Secondary: 53C35.

 Citation:

• Figure 1.  The Pseudo-Code For SFL

Figure 2.  This is Table 1

Figure 3.  This is table2

Figure 4.  This is table3

Figure 5.  This is table4

Figure 6.  This is table5

Figure 7.  This is table6

Figure 8.  This is table7

Figure 9.  This is table8

Figure 10.  This is table9

Figure 11.  This is table10

Figure 12.  This is table11

Figure 13.  This is table12

•  [1] A. Asuncion and D. Newman, Uci machine learning repository, 2007. [2] G. A. Carpenter, S. Grossberg, N. Markuzon, J. H. Reynolds and D. B. Rosen, Fuzzy artmap: A neural network architecture for incremental supervised learning of analog multidimensional maps, IEEE Transactions on Neural Networks, 3 (1992), 698-713.  doi: 10.1109/72.159059. [3] G. A. Carpenter, S. Grossberg and J. H. Reynolds, ARTMAP: Supervised Real-Time Learning and Classification of Nonstationary Data by a Self-Organizing Neural Network Elsevier Science Ltd. , 1991. doi: 10.1109/ICNN.1991.163370. [4] N. V. Chawla, N. Japkowicz and A. Kotcz, Editorial: Special issue on learning from imbalanced data sets, Acm Sigkdd Explorations Newsletter, 6 (2004), 1-6. [5] G. Ditzler, M. D. Muhlbaier and R. Polikar, Incremental learning of new classes in unbalanced datasets: Learn?+?+?.UDNC, International Workshop on Multiple Classifier Systems, Multiple Classifier Systems, (2010), 33-42.  doi: 10.1007/978-3-642-12127-2_4. [6] G. Ditzler, R. Polikar and N. Chawla, An incremental learning algorithm for non-stationary environments and class imbalance, International Conference on Pattern Recognition, (2010), 2997-3000.  doi: 10.1109/ICPR.2010.734. [7] Y. Freund and R. E. Schapire, A short introduction to boosting, Journal of Japanese Society for Artificial Intelligence, 14 (1999), 771-780. [8] L. Fu, H.-H. Hsu and J. C. Principe, Incremental backpropagation learning networks, IEEE Transactions on Neural Networks, 7 (1996), 757-761. [9] H. He and E. A. Garcia, Learning from imbalanced data, IEEE Transactions on Knowledge and Data Engineering, 21 (2009), 1263-1284. [10] H. Inoue and H. Narihisa, Self-organizing neural grove and its applications, IEEE International Joint Conference on Neural Networks, 2 (2005), 1205-1210.  doi: 10.1109/IJCNN.2005.1556025. [11] N. Japkowicz and S. Stephen, The Class Imbalance Problem: A Systematic Study IOS Press, 2002. [12] N. Kasabov, Evolving fuzzy neural networks for supervised/unsupervised online knowledge-based learning, IEEE Transactions on Systems Man & Cybernetics Part B Cybernetics A Publication of the IEEE Systems Man & Cybernetics Society, 31 (2001), 902-918.  doi: 10.1109/3477.969494. [13] M. Lin, K. Tang and X. Yao, Dynamic sampling approach to training neural networks for multiclass imbalance classification, IEEE Transactions on Neural Networks and Learning Systems, 24 (2013), 647-660. [14] Y. Liu and X. Yao, Simultaneous training of negatively correlated neural networks in an ensemble, IEEE Transactions on Systems Man & Cybernetics Part B Cybernetics A Publication of the IEEE Systems Man & Cybernetics Society, 29 (1999), 716-725. [15] F. L. Minku, H. Inoue and X. Yao, Negative correlation in incremental learning, Natural Computing, 8 (2009), 289-320.  doi: 10.1007/s11047-007-9063-7. [16] M. Muhlbaier, A. Topalis and R. Polikar, Incremental learning from unbalanced data, In Neural Networks, 2004. Proceedings. 2004 IEEE International Joint Conference on, IEEE, 2 (2004), 1057-1062.  doi: 10.1109/IJCNN.2004.1380080. [17] M. Muhlbaier, A. Topalis and R. Polikar, Learn++.mt: A new approach to incremental learning, Lecture Notes in Computer Science, 3077 (2004), 52-61.  doi: 10.1007/978-3-540-25966-4_5. [18] M. D. Muhlbaier, A. Topalis and R. Polikar, Learn ++. nc: combining ensemble of classifiers with dynamically weighted consult-and-vote for efficient incremental learning of new classes, IEEE Transactions on Neural Networks 20 (2009), p152. [19] S. Ozawa, S. Pang and N. Kasabov, Incremental learning of chunk data for online pattern classification systems, IEEE Trans Neural Netw, 19 (2008), 1061-1074.  doi: 10.1109/TNN.2007.2000059. [20] R. Polikar, J. Byorick, S. Krause and A. Marino, Learn++: A classifier independent incremental learning algorithm for supervised neural networks, International Joint Conference on Neural Networks, (2002), 1742-1747.  doi: 10.1109/IJCNN.2002.1007781. [21] R. Polikar, L. Upda, S. S. Upda and V. Honavar, Learn++: an incremental learning algorithm for supervised neural networks, IEEE Transactions on Systems Man & Cybernetics Part C, 31 (2001), 497-508.  doi: 10.1109/5326.983933. [22] M. Salganicoff, Tolerating concept and sampling shift in lazy learning using prediction error context switching, Artificial Intelligence Review, 11 (1997), 133-155.  doi: 10.1007/978-94-017-2053-3_5. [23] M. C. Su, J. Lee and K. L. Hsieh, A new artmap-based neural network for incremental learning, Neurocomputing, 69 (2006), 2284-2300.  doi: 10.1016/j.neucom.2005.06.020. [24] Y. Sun, M. S. Kamel and Y. Wang, Boosting for learning multiple classes with imbalanced class distribution, In Data Mining, 2006. ICDM'06. Sixth International Conference on, IEEE, (2006), 592-602.  doi: 10.1109/ICDM.2006.29. [25] E. K. Tang, P. N. Suganthan and X. Yao, An analysis of diversity measures, Machine Learning, 65 (2006), 247-271.  doi: 10.1007/s10994-006-9449-2. [26] K. Tang, M. Lin, F. L. Minku and X. Yao, Selective negative correlation learning approach to incremental learning, Neurocomputing, 72 (2009), 2796-2805.  doi: 10.1016/j.neucom.2008.09.022. [27] W. X. Wen, H. Liu and A. Jennings, Self-generating neural networks, International Joint Conference on Neural Networks, 4 (2002), 850-855. [28] G. Widmer and M. Kubat, Effective learning in dynamic environments by explicit context tracking, In Machine learning: ECML-93, Springer, 667 (1993), 227-243.  doi: 10.1007/3-540-56602-3_139. [29] J. R. Williamson, Gaussian artmap: A neural network for fast incremental learning of noisy multidimensional maps, Neural Networks, 9 (1996), 881-897.  doi: 10.1016/0893-6080(95)00115-8.

Figures(13)