Time series prediction with improved neuro-endocrine model

Abstract

The paper is focused on improving the performance of neuro-endocrine models with considering the interaction of glands. Comparing to conventional neuro-endocrine models, the concentration of hormone of one gland is modulated by those of others, and the weights of cells are modulated by the improved endocrine system. The interacted equation among all glands is designed and the parameters of them are chosen with theory analysis. Because all the parameters of the model are constants when the system reaches the equilibrium state, particle swarm optimization algorithm is utilized to search the optimal parameters of the model. The theory analysis indicates that the performance of neuro-endocrine model is better than or at least equal to that of corresponding artificial neural network. To indicate the effectiveness of the proposed model, some time series from different research fields, which are used in some literatures, are tested with the proposed model, the results indicate that the proposed model has some good performance.

GA-based learning for rule identification in fuzzy neural networks

“…The model error presented in Table 2 for the GAFNN is for this unseen test data set. The improved Neuro-Endocrine model proposed in [37] uses the interaction mechanism of different glands for regulating the neural network. The model improved the predictive accuracy of time series.…”

Prediction of Electroencephalogram Time Series With Electro-Search Optimization Algorithm Trained Adaptive Neuro-Fuzzy Inference System

“…In another work, Forney [57] used Elman RNN to classify EEG time series by including use of Winner-Takes-All, Linear Discriminant Analysis, and Quadratic Discriminant Analysis as well as alternative prediction stages. Chen et al [58] used an improved Neuro-Endocrine Model (INEM), which was supported by Linearly Decreasing Weight Particle Swarm Optimization (LDWPSO) for predicting some chaotic time series including also EEG data. Obtained results reveal well-enough performances by the INEM-LDWPSO system developed by the authors.…”

Modeling of Dynamic Systems Using Orthogonal Endocrine Adaptive Neuro-Fuzzy Inference Systems

“…Idea is to mimic the process from the nature where the behavior of the neurons is regulated by hormones secreted by glands of endocrine system. In contrast to these papers [6][7][8] where hormones are usually used to modify weights of neurons in artificial neural networks, based on external conditions, idea of this paper is to introduce these stimuli directly to the neurons of fourth layer, via parameter d in Eq. (11), as reaction to the variations of nominal components of dynamical system under effect of changed working conditions, determined by sensors.…”

“…These networks mimic the biological process of regulating the behavior of the system through hormone secretion from the gland cell associated to the artificial neural network. This principle of adding hormones can be applied to all kinds of regular neural networks [7,8] as well as ANFIS, as in this paper. The advantage of the method presented in this paper is that the system after completion of the training has the ability to adapt to ever changing conditions in which the system operates, based on the information obtained from the external sensors.…”