Genetic Programming-Based Chaotic Time Series Modeling

Wei Zhang,Zhi-ming Wu,Gen-ke Yang
DOI: https://doi.org/10.1631/jzus.2004.1432
2004-01-01
Journal of Zhejiang University SCIENCE A
Abstract:This paper proposes a Genetic Programming-Based Modeling (GPM) algorithm on chaotic time series. GP is used here to search for appropriate model structures in function space, and the Particle Swarm Optimization (PSO) algorithm is used for Nonlinear Parameter Estimation (NPE) of dynamic model structures. In addition, GPM integrates the results of Nonlinear Time Series Analysis (NTSA) to adjust the parameters and takes them as the criteria of established models. Experiments showed the effectiveness of such improvements on chaotic time series modeling.
What problem does this paper attempt to address?