The use of Recurrent Neural Networks is not as extensive as Feedforward Neural Networks. Training algorithms for Recurrent Neural Networks, based on the error gradient, are very unstable in their search for a minimum and require much computational time when the number of neurons is high. The problems surrounding the application of these methods have driven us to develop new training tools. In t...

This paper proposes a recurrent wavelet-based neuro-fuzzy system (RWNFS) with a reinforcement group cooperation-based symbiotic evolution (R-GCSE) for solving various control problems. The R-GCSE is different from the traditional symbiotic evolution. In the R-GCSE method, a population is divided to several groups. Each group formed by a set of chromosomes represents a fuzzy rule and compensatio...

A Recurrent Trainable Neural Network (RTNN) with a two layer canonical architecture learned by a dynamic Backpropagation learning algorithm is incorporated in a Hierarchical Fuzzy-Neural Multi-Model (HFNMM) identifier, combining the fuzzy model flexibility with the learning abilities of the RTNNs. The local and global features of the proposed HFNMM identifier are implemented by a Hierarchical S...

The paper proposed to use recurrent Fuzzy-Neural Multi-Model (FNMM) identifier for decentralized identification of a distributed parameter anaerobic wastewater treatment digestion bioprocess, carried out in a fixed bed and a recirculation tank. The distributed parameter analytical model of the digestion bioprocess is used as a plant data generator. It is reduced to a lumped system using the ort...

This paper presents an innovative hybrid approach for online recognition of handwritten symbols. This approach is composed of two main techniques. The first technique, based on fuzzy logic, deals with feature extraction from a handwritten stroke and the second technique, a recurrent neural network, uses the features as an input to recognise the symbol. In this paper we mainly focuss our study o...

In this thesis we present two new methods for discovering knowledge from ordered datasets using Baldwin’s mass assignment and the Fril programming language. We derive our methods from human observation and knowledge handling computing techniques. We define this combination of observed human behaviour and conventional computing as humanist computing. The first method represents trends in ordered...

In this article the author presents JFK, which stands for Java Fuzzy Kit. JFK is an Application Programming Interface (API) that complies with both, a general structure of a fuzzy rule base and the necessary processing to compute the generalized principle of extension. A recurrent structure is found for a class of fuzzy expert systems, known as the Mamdani model. This leads to claim that a desi...

Dynamical Recurrent Neural Networks (DRNN) (Aussem 1995a) are a class of fully recurrent networks obtained by modeling synapses as autoregressive filters. By virtue of their internal dynamic, these networks approximate the underlying law governing the time series by a system of nonlinear difference equations of internal variables. They therefore provide history-sensitive forecasts without havin...