Абстрактный

Theory and algorithm of state space model and its application in financial econometrics

Yaoyuan Liu

Because the time series of financial econometrics has feature of nonlinearity and is nongaussian as well as the latent variable and time-varying parameters. The standard quantitative analysis model in the practice is very difficult. This thesis manages to base on the theory of recurrence Bayes filtering theory and integrate the state space model of modern statistical calculation. The thesis studies the theory and algorithm of state space model, and through the comparison and check in the simulation experiment we find that the application of state space model in financial econometrics is the best.