Impact of fuzzy ARTMAP match tracking strategies on the recognition of handwritten digits

Applying fuzzy ARTMAP to complex real-world problems such as handwritten character recognition may lead to poor performance and a convergence problem whenever the training set contains very similar or identical patterns that belong to different classes. To circumvent this problem, some alternatives to the network’s original match tracking (MT) process have been proposed in literature, such as using negative MT, and removing MT altogether. In this paper, the impact on fuzzy ARTMAP performance of different MT strategies is assessed using different patterns recognition problems – two types of synthetic data as well as a real-world handwritten digit data. Fuzzy ARTMAP is trained with the original match tracking (MT+), with negative match tracking (MT-), and without MT algorithm (WMT), and the performance of these networks is compared to the case where the MT parameter is optimized using a PSO-based strategy, denoted PSO(MT). Through a comprehensive set of computer simulations, it has been observed that by training with MT-, fuzzy ARTMAP expends fewer resources than other MT strategies, but can achieve a significantly higher generalization error, especially for data with overlapping class distributions. Generalization error achieved using WMT is significantly higher than other strategies on non overlapping data with complex non-linear decision bounds. Furthermore, the number of internal categories increases significantly. PSO(MT) yields the best overall generalization error, and a number of internal categories that is smaller than WMT, but generally higher than MT+ and MT-. However, this strategy requires a large number of training epochs to convergence.

[1]  M. Stone.  Cross‐Validatory Choice and Assessment of Statistical Predictions , 1976 .

[2]  Mohamed Cheriet,et al.  Estimating accurate multi-class probabilities with support vector machines , 2005, Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005..

[3]  R. Sabourin,et al.  Factors of overtraining with fuzzy ARTMAP neural networks , 2005, Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005..

[4]  Stephen Grossberg,et al.  ARTMAP: supervised real-time learning and classification of nonstationary data by a self-organizing neural network , 1991, [1991 Proceedings] IEEE Conference on Neural Networks for Ocean Engineering.

[5]  Gail A. Carpenter,et al.  ARTMAP-IC and medical diagnosis: Instance counting and inconsistent cases , 1998, Neural Networks.

[6]  Georgios C. Anagnostopoulos,et al.  Putting the Utility of Match Tracking in Fuzzy ARTMAP Training to the Test , 2003, KES.

[7]  Robert F. Harrison,et al.  A modified fuzzy ARTMAP architecture for the approximation of noisy mappings , 1995, Neural Networks.

[8]  Guszti Bartfai,et al.  An improved learning algorithm for the fuzzy ARTMAP neural network , 1995, Proceedings 1995 Second New Zealand International Two-Stream Conference on Artificial Neural Networks and Expert Systems.

[9]  Luiz Eduardo Soares de Oliveira,et al.  Automatic Recognition of Handwritten Numerical Strings: A Recognition and Verification Strategy , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[10]  Stephen Grossberg,et al.  Fuzzy ARTMAP: A neural network architecture for incremental supervised learning of analog multidimensional maps , 1992, IEEE Trans. Neural Networks.

[11]  Luiz Eduardo Soares de Oliveira,et al.  Particle Swarm Optimization of Fuzzy ARTMAP Parameters , 2006, The 2006 IEEE International Joint Conference on Neural Network Proceedings.