Search

Fuzzy ART and Fuzzy ARTMAP models arise from the synergy between the Fuzzy Set Theory and the Adaptive Resonance paradigm (ART). In this work, the performance of these models and the use of Fuzzy ARTMAP for function approximation are studied. In a first analysis, a relationship between the model parameters and the features of the generated categories is established. In the second part, the connection between these categories and the capacity of prediction of the model is analytically described. Joining these two studies, the link between the parameters and the prediction error of the model is found, in the form of bounds for the prediction error depending on the model parameters and the characteristics of the data used in the learning. These results provide a quantitative description of the parameter influence on the architecture behavior, opening the use of Fuzzy ARTMAP as a model for the unknown dynamic system identification from input/output data. To illustrate the theoretical developments, several experiments have been carried out using different kinds of functions, which show the accuracy of the proposed bounds. [ABSTRACT FROM AUTHOR]

Abstract: In this work, local stability on the initialization phase of nonlinear autoregressive with exogenous inputs multilayer perceptrons (NARX MLP) and radial basis functions (NARX RBF) neural networks is studied. It will be shown that the selection of adequate ranges for the initial weights is related with local stability of the network in its initial stage. As a result, quantitative limits for the initial weights are established that guarantee local stability and accelerate the learning process. These theoretical developments have been tested in experiments which corroborate the improvements achieved with the proposed initialization methods. [Copyright &y& Elsevier]

This work develops and tests a neighborhood-based approach to the Gauss-Newton Bayesian regularization trainingmethod for feedforward backpropagation networks. The proposed method improves the training efficiency, significantly reducing requirements on memory and computational time while maintaining the good generalization feature of the original algorithm. This version of the Gauss-Newton Bayesian regularization greatly expands the scope of application of the original method, as it allows training networks up to 100 times larger without losing performance. [ABSTRACT FROM AUTHOR]

A neuro-fuzzy system for isolated hand-written digit recognition using a similarity fuzzy measure is presented. The system is composed of two main blocks: a first block that normalises the input and compares it with a set of fuzzy patterns, and a second block with a multilayer perceptron (MLP) to perform the definitive classification. The comparison with the fuzzy patterns is carried out via a fuzzy similarity measure that uses the Yager parametric norms and co-norms. Along this work, several values of the parameter have been studied, in order to obtain the optimum. The simplicity of the method makes it extremely quick. Recognition accuracy of the method is about 90% single classification, and close to 97,5% classification scheme. [ABSTRACT FROM AUTHOR]

The temporal character of the input is, generally, not taken into account in the neural models. This paper presents an extension of the FasArt model focused on the treatment of temporal signals. FasArt model is proposed as an integration of the characteristic elements of the Fuzzy System Theory in an ART architecture. A duality between the activation concept and membership function is established. FasArt maintains the structure of the Fuzzy ARTMAP architecture, implying a static character since the dynamic response of the input is not considered. The proposed novel model, dynamic FasArt (dFasArt), uses dynamic equations for the processing stages of FasArt: activation, matching and learning. The new formulation of dFasArt includes time as another characteristic of the input. This allows the activation of the units to have a history-dependent character instead of being only a function of the last input value. Therefore, dFasArt model is robust to spurious values and noisy inputs. As experimental work, some cases have been used to check the robustness of dFasArt. A possible application has been proposed for the detection of variations in the system dynamics. [Copyright &y& Elsevier]

Abstract: The arriving order of data is one of the intrinsic properties of a signal. Therefore, techniques dealing with this temporal relation are required for identification and signal processing tasks. To perform a classification of the signal according with its temporal characteristics, it would be useful to find a feature vector in which the temporal attributes were embedded. The correlation and power density spectrum functions are suitable tools to manage this issue. These functions are usually defined with statistical formulation. On the other hand, in biology there can be found numerous processes in which signals are processed to give a feature vector; for example, the processing of sound by the auditory system. In this work, the dNSP (dynamic Neural Signal Processing) architecture is proposed. This architecture allows representing a time-varying signal by a spatial (thus statical) vector. Inspired by the aforementioned biological processes, the dNSP performs frequency decomposition using an analogical parallel algorithm carried out by simple processing units. The architecture has been developed under the paradigm of a multilayer neural network, where the different layers are composed by units whose activation functions have been extracted from the theory of Neural Dynamic [Grossberg, S. (1988). Nonlinear neural networks principles, mechanisms and architectures. Neural Networks, 1, 17–61]. A theoretical study of the behavior of the dynamic equations of the units and their relationship with some statistical functions allows establishing a parallelism between the unit activations and correlation and power density spectrum functions. To test the capabilities of the proposed approach, several testbeds have been employed, i.e. the frequencial study of mathematical functions. As a possible application of the architecture, a highly interesting problem in the field of automatic control is addressed: the recognition of a controlled DC motor operating state. [Copyright &y& Elsevier]