Using genetic algorithms to parameters (d,r) estimation for threshold autoregressive models

Authors:
Berlin Wu;Chih-Li Chang
Affiliations:
Department of Mathematical Sciences, National Chengchi University, Taiwan;Department of Mathematics, Hsiuping Institute of Technology, Taiwan
Venue:
Computational Statistics & Data Analysis
Year:
2002

Citing 5
Cited 1

Model-free forecasting for nonlinear time series (with application to exchange rates)

Computational Statistics & Data Analysis
Fuzzy connectives based crossover operators to model genetic algorithms population diversity

Fuzzy Sets and Systems
Extracting fuzzy rules for system modeling using a hybrid of genetic algorithms and Kalman filter

Fuzzy Sets and Systems
Use of fuzzy statistical technique in change periods detection of nonlinear time series

Applied Mathematics and Computation
Nonlinear parameter estimation via the genetic algorithm

IEEE Transactions on Signal Processing

Annealing evolutionary stochastic approximation Monte Carlo for global optimization

Statistics and Computing

Quantified Score

Hi-index	0.03

Visualization

Abstract

Threshold autoregressive model (TAR model) has certain characteristics due to which linear models fail to fit a nonlinear time series, while the problem of how to find an appropriate threshold value still attracts many researchers' attention. In this paper, we apply the genetic algorithms to estimate the threshold and lag parameters r and d for TAR models. The selection operator is formulated following Darwin's principle of survival of the fittest to guide the trek through a search space. The crossover and mutation operators have been inspired by the mechanisms of gene mutation and chromosome recombination.