Time series prediction by artificial neural networks and differential evolution in distributed environment

Authors:
Todor Balabanov;Iliyan Zankinski;Nina Dobrinkova
Affiliations:
Institute of Information and Communication Technologies, Bulgarian Academy of Science, Sofia, Bulgaria;Faculty of Computer Systems and Management, Technical University of Sofia, Sofia, Bulgaria;Institute of Information and Communication Technologies, Bulgarian Academy of Science, Sofia, Bulgaria
Venue:
LSSC'11 Proceedings of the 8th international conference on Large-Scale Scientific Computing
Year:
2011

Citing 3
Cited 0

Neural Networks: A Comprehensive Foundation

Neural Networks: A Comprehensive Foundation
Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces

Journal of Global Optimization
Noisy Time Series Prediction using Recurrent Neural Networks and Grammatical Inference

Machine Learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

Current work will present a model for time series prediction by the usage of Artificial Neural Networks (ANN) trained with Differential Evolution (DE) in distributed computational environment. Time series prediction is a complex work and demand development of more effective and faster algorithms. ANN is used as a base and it is trained with historical data. One of the main problems is how to select accurate ANN training algorithm. There are two general possibilities -- exact numeric optimization methods and heuristic optimization methods. When the right heuristic is applied the training can be done in distributed computational environment. In this case there is much faster and realistic output, which helps to achieve better prediction.