USP Electronic Research Repository

Modified Neuron-Synapse level problem decomposition method for Cooperative Coevolution of Feedforward Networks for Time Series Prediction

Nand, Ravneil and Sharma, Bibhya N. (2017) Modified Neuron-Synapse level problem decomposition method for Cooperative Coevolution of Feedforward Networks for Time Series Prediction. [Conference Proceedings]

[thumbnail of Modified_Neuron-Synapse_level_problem_decomposition_method_for_Cooperative_Coevolution_of_Feedforward_Networks_for_Time_Series_Prediction.pdf] PDF - Accepted Version
Restricted to Registered users only

Download (203kB) | Request a copy

Abstract

Complex problems have been solved efficiently through decomposition of a particular problem using problem decompositions. Even combination of different distinct problem decomposition methods has shown good results in time series prediction. The application of different problem decomposition methods at different stages of a network can share its strengths to solve the problem in hand better. Hybrid versions of two distinct problem decomposition methods has showed promising results in past. In this paper, a modified version of latterly introduced Neuron-Synapse level problem decomposition is proposed using feedforward neural networks for time series prediction. The results shows that the proposed modified model has got better results in more datasets when compared to its previous version. The results are better in some cases for proposed method in comparison to several other methods from the literature.

Item Type: Conference Proceedings
Additional Information: Accepted for publishing
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Faculty of Science, Technology and Environment (FSTE) > School of Computing, Information and Mathematical Sciences
Depositing User: Fulori Nainoca - Waqairagata
Date Deposited: 21 Mar 2018 00:02
Last Modified: 10 Jun 2018 21:54
URI: https://repository.usp.ac.fj/id/eprint/10622

Actions (login required)

View Item View Item