Approximate Parameter Estimation for Mixtures of Vector Autoregressive Stationary Time Series Models


Христов, Любомир (2007) Approximate Parameter Estimation for Mixtures of Vector Autoregressive Stationary Time Series Models Internaional Conference on Computer Systems and Technlologies – CompSysTech'2007, Rousse, June 2007


 Here we consider a mixture of autoregressive time series models. This mixture can be considered as a peculiar simplification of the classical autoregressive model with changes in regime introduced by Hamilton [1]. Here the mathematics is easier than in the classical case based on the Hidden Markov Model technology therefore the calculations are more quick and stable. Our numerical experiments show almost perfect case identification for the simulated data.
  Статия
 Vector Autoregressive Model, Changes in Regime


Природни науки, математика и информатика

Natural sciences, mathematics and informatics

 Издадено
  4725
 Любомир Христов

Научният архив поддържа инициативата за отворен достъп OAI 2.0 с начален адрес: http://da.uni-vt.bg/oai2/