一种基于粒子群算法改进的ARCH模型及对国内油价波动的研究

The Improved ARCH Model Based on Particle Swarm Optimization Algorithm and the Study of Domestic Oil Price Fluctuations

传统的ARCH模型族在参数估计中使用的极大似然估计存在鲁棒性差和易收敛到局部最优解的缺点,为克服传统参数估计的上述缺点,提出了基于粒子群算法改进的智能算法,并利用粒子群算法对国内油价建立了AR(1)-ARCH、AR(1)-TARCH(1)模型.

Traditional ARCH model used maximum likelihood estimation to estimate parameter which exist low robustness and convergence to local optimal solution.In order to overcome the shortcomings of traditional parameter estimation,this paper puts an intelligent algorithm based on Particle Swarm Optimization（PSO）algorithm and builds ARCH models of domestic oil prices.