ABSTRACT
We develop a model to forecast Chinese soybean futures price with eighteen predictors by integrating the recently proposed dynamic model averaging (DMA) and particle swarm optimization (PSO). Specifically, three important parameters, i.e., two forgetting factors and a decay factor, of DMA are tuned by PSO. The proposed prediction model, named DMA-PSO, not only allow for coefficients to change over time, but also allow for forecasting model to evolve over time. Experimental results show that the proposed DMA-PSO outperforms four counterparts and the best predictors in DMA-PSO for forecasting soybean futures price vary a lot over time1.
- A. E. Raftery, M. Kárný and P. Ettler, 2010. Online prediction under model uncertainty via dynamic model averaging: Application to a cold rolling mill. Technometrics 52(1), pp. 52--66.Google ScholarCross Ref
- J. Kennedy, R. Eberhart, 1995, Particle swarm optimization, in: Proceedings of the 1995 IEEE International Conference on Neural Networks, pp. 1942--1948.Google ScholarCross Ref
Recommendations
Euclidean Particle Swarm Optimization
ICINIS '09: Proceedings of the 2009 Second International Conference on Intelligent Networks and Intelligent SystemsParticle swarm optimization (PSO) is a swarm intelligence algorithm, has been successfully applied to many engineering optimization problems and shown its high search speed in these applications. However, as the dimension and the number of local optima ...
Dynamic Multi-swarm Particle Swarm Optimization with Center Learning Strategy
Advances in Swarm IntelligenceAbstractIn this paper, we propose a novel variant of particle swarm optimization, called dynamic multi-swarm particle swarm optimization with center learning strategy (DMPSOC). In DMPSOC, all particles are divided into several sub-swarms. Then, a center-...
Mechanism of Particle Swarm Optimization and Analysis on Its Convergence
ISIP '10: Proceedings of the 2010 Third International Symposium on Information ProcessingParticle swarm optimization (PSO) is a new swarm intelligence algorithm, derived from artificial life and evolutionary computation theory. It makes full use of the information--sharing particles of the cluster to obtain the optimal solution of the ...
Comments