Intelligent forecasting of S&P 500 time series: a self-organizing fuzzy approach

Authors:
Chunshien Li;Hsin Hui Cheng
Affiliations:
Laboratory of Intelligent Systems and Applications, Department of Information Management, National Central University, Taiwan, R.O.C;Laboratory of Intelligent Systems and Applications, Department of Information Management, National Central University, Taiwan, R.O.C
Venue:
ACIIDS'11 Proceedings of the Third international conference on Intelligent information and database systems - Volume Part II
Year:
2011

Citing 7
Cited 0

Neuro-fuzzy and soft computing: a computational approach to learning and machine intelligence

Neuro-fuzzy and soft computing: a computational approach to learning and machine intelligence
Recent approaches to global optimization problems through Particle Swarm Optimization

Natural Computing: an international journal
Recurrent neuro-fuzzy hybrid-learning approach to accurate system modeling

Fuzzy Sets and Systems
Soft-computing techniques and ARMA model for time series prediction

Neurocomputing
PSO-based single multiplicative neuron model for time series prediction

Expert Systems with Applications: An International Journal
Complex neuro-fuzzy self-learning approach to function approximation

ACIIDS'10 Proceedings of the Second international conference on Intelligent information and database systems: Part II
Self-organizing neuro-fuzzy system for control of unknown plants

IEEE Transactions on Fuzzy Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Stock index time series may allow investors to become aware of the change of stock market. In the paper, we aim at forecasting S&P 500 Index, one of the most representative stock indices in United States. A self-organizing fuzzy-based approach for intelligent predictor is used. The design for the predictor is divided into the structure and parameter learning stages. The FCMBased Splitting Algorithm is used to determine the optimal number of fuzzy rules for the predictor. Two hybrid learning algorithms, the PSO-RLSE and PSO-RLSE-PSO methods, are used for the parameter learning of the predictor, respectively. To test the proposed approach, we devise experiments to compare the performances by the intelligent predictor trained with the two learning algorithms, respectively. Moreover, an additional experiment for different input orders is conducted to see the influence on the performance. The excellent performances in accuracy by the proposed intelligent approach are exposed.