Martha Pulido

Particle swarm optimization of ensemble neural networks with fuzzy aggregation for time series prediction of the Mexican Stock Exchange

Abstract This paper describes a hybrid method based on particle swarm optimization for designing ensemble neural networks with fuzzy aggregation of responses to forecast complex time series. The time series that was considered in this paper, to compare the hybrid approach with traditional methods, is the Mexican Stock Exchange, and the results shown are for the optimization of the structure of the ensemble neural network with type-1 and type-2 fuzzy logic integration. Simulation results show that the optimized ensemble neural network approach produces good prediction of the Mexican Stock Exchange.

Comparison of fuzzy controllers for the water tank with Type-1 and Type-2 fuzzy logic

In this paper a comparison of the use of Type-1 and Type-2 fuzzy logic in the benchmark problem known as the problem of the water tank is presented. Fuzzy logic allows managing uncertainty, which is very common in linguistic fuzzy systems. It has been shown that Type-2 fuzzy systems manage better the uncertainty in real world problems but, evidence exists also to show that using Type-2 technology is usually computationally more expensive. It is shown in this paper using an experimental basis that an interval Type-2 Fuzzy System presents better results than those obtained with the traditional way of handling fuzzy systems with Type-1 fuzzy logic. Experiments and comparisons are presented to validate the proposed approach.

Genetic optimization of ensemble neural networks for complex time series prediction

This paper describes an optimization method for ensemble neural network models with fuzzy aggregation of responses for forecasting complex time series using genetic algorithms. The time series under consideration for testing the hybrid approach is the Mackey-Glass data, and results for the optimization of type-1 fuzzy response aggregation in the ensemble neural network are presented. Simulation results show the effectiveness of the proposed approach.

An Ensemble Neural Network Architecture with Fuzzy Response Integration for Complex Time Series Prediction

In this paper we describe the application of an architecture for an ensemble neural network for Complex Time Series Prediction. The times series we are considering are: the Mackey-Glass, Dow Jones and Mexican Stock Exchange and we show the results of a set of trainings with the ensemble neural network, and its integration with the methods of average, weighted average and Fuzzy Integration. Simulation results show very good prediction of the ensemble neural network with fuzzy logic integration.

Optimization of type-2 fuzzy integration in ensemble neural networks for predicting the US Dolar/MX pesos time series

This paper describes the optimization of an ensemble neural network with fuzzy integration of responses based on type-1 and type-2 fuzzy logic. Genetic algorithms are used as method of optimization in this case. The time series that is being considered for the ensemble is the US Dollar/MX Peso exchange rate. Simulation results show that the ensemble approach produces good prediction of the exchange rate US Dollar/MX Peso.

A New Method for Type-2 Fuzzy Integration in Ensemble Neural Networks Based on Genetic Algorithms

This paper describes a proposed method for type-2 fuzzy integration that can be used in the fusion of responses for an ensemble neural network. We consider the case of the design of a type-2 fuzzy integrator for fusion of a neural network ensemble. The network structure of the ensemble may have a maximum of 5 modules. This integrator consists of 32 fuzzy rules, with 5 inputs depending on the number of modules of the neural network ensemble and one output. Each input and output linguistic variable of the fuzzy system uses Gaussian membership functions. The performance of type-2 fuzzy integrators is analyzed under different levels of uncertainty to find out the best design of the membership functions. In this case the proposed method is applied to time series prediction.

Optimization of type-2 fuzzy integration in ensemble neural networks for predicting the Dow Jones time series

This paper describes an optimization method based on genetic algorithms for ensemble neural networks with type-2 fuzzy integration with application to the forecasting of complex time series. The time series that was considered in this paper, to compare the hybrid genetic-neuro-fuzzy approach with traditional methods is the Dow Jones, and the results shown are for the optimization of the structure of the ensemble neural network and type-2 fuzzy integration. Simulation results show that the ensemble approach produces good prediction of the Dow Jones time series.

Ensemble neural networks with fuzzy logic integration for complex time series prediction

In this paper the application of an ensemble neural network model for complex time series prediction is presented. The Mackey-Glass time series is considered for testing the ensemble model. Simulation results of the ensemble neural network and its integration with the methods of average, weighted average and fuzzy integration are presented. Simulation results show very good prediction of the ensemble neural network with the method of fuzzy integration.

Optimization of ensemble neural networks with type-2 fuzzy response integration for predicting the Mackey-Glass time series

This paper describes the optimization of an ensemble neural network with fuzzy integration of responses based on type-1 and type-2 fuzzy logic. Genetic algorithms are used as a method of optimization for the ensemble model in this case of study. The time series that is being considered is the Mackey-Glass benchmark. Simulation results show that the ensemble approach produces good prediction of the Mackey-Glass time series.

Optimization of ensemble neural networks with fuzzy integration using the particle swarm algorithm for the US Dollar/MX time series prediction

This paper describes the design with Particle Swarm Optimization of a neural network ensemble with type-1 and type-2 fuzzy integration of responses. The proposed ensemble neural network approach is tested with the problem of time series prediction. The time series that is being considered for testing the hybrid approach is the US/Dollar MX time series. Simulation results show that the ensemble neural network approach produces good prediction of the Dollar time series.