Advanced Developments and Applications of the Fuzzy ARTMAP Neural Network in Pattern Classification

Abstract

Abstract. Since its inception in 1992, the fuzzy ARTMAP (FAM) neural network (NN) has attracted researchers' attention as a fast, accurate, off and online pattern classifier. Since then, many studies have explored different issues concerning FAM optimization, training and evaluation, e.g., model sensitivity to parameters, ordering strategy for the presentation of the training patterns, training method and method of predicting the classification accuracy. Other studies have suggested variants to FAM to improve its generalization capability or overcome the prime limitation of the model, which is category proliferation (i.e., model complexity that increases with data complexity). Category proliferation is pronounced in problems that are noisy or contain a large degree of class statistical overlapping. In many investigations, FAM was improved by incorporating elements of optimization theory, Bayes' decision theory, evolutionary learning, and cluster analysis. Due to its appealing characteristics, FAM and its variants have been applied extensively and successfully to real-world classification problems. Numerous applications were reported in, for example, the processing of signals from different sources, images, speech, and text; recognition of speakers, image objects, handwritten, and genetic abnormalities; and medical and fault diagnoses. When compared to other state-of-the-art machine learning classifiers, FAM and its variants showed superior speed and ease of training, and in most cases they delivered comparable classification accuracy.

Signal classification based on spectral redundancy and neural network ensembles

“…This type of learning strategy is based on the construction of a nearest neighbor look-up table [10].A SFAm performes a classification of a pattern whose complement coded representation is denoted by I, by the following three steps: prototype choice, prototype match, and learning [7].…”

“…Since SFAM methods are sensitive to the order of the training patterns [7], [8], our discussion will be limited to bagging and boosting. In fact, these methods rely on resampling techniques to obtain different training sets for each classifier to achieve diversity.…”

“…The success of a CR system used as a spectral occupancy detector can be intimately related to the accuracy, robustness, and fast convergence of the selected classifier. Most state-ofthe-art pattern classifiers methods usually need long training time and large sample sizes to achieve good performances [7]. These requirements make the training process computationally expensive, an undesiderable feature in a CR system that can operate in an unknown environment and can perform an online training.…”

“…Although it has been reported that SFAM and other fuzzy ARTMAP versions have accurate and fast learning in performing various classification tasks [7], their major drawback is that the network performances are affected by the ordering and by the sequence of training examples [8]. Moreover, in [7] has reported that SFAM networks are sensitive to statistical overlapping between the classes.…”

“…Moreover, in [7] has reported that SFAM networks are sensitive to statistical overlapping between the classes.…”