Research On Parameter Estimation Of SV Model Based On Approximate Bayesian Computation Method

In the financial market,there is a general fluctuation phenomenon in the financial time series data.In a large number of studies on volatility,the stochastic volatility(SV)model has been proved to be able to better describe the volatility characteristics of the financial time series with time,accurately calculate the quantitative characteristics of volatility,and describe the volatility more widely and flexibly.It is not easy to estimate the parameters of the volatility model by using the maximum likelihood method,which makes it difficult to estimate the parameters of the volatility model,being able to estimate the parameters of the stochastic volatility model quickly and accurately will help to expand the application space of the stochastic volatility model in the financial field,which has a certain practical significance.Firstly,the common parameter estimation methods of SV model are sorted out,and the most commonly used Markov chain Monte Carlo(MCMC)method is further discussed and studied.Secondly,through the tools of approximate Bayesian theory and conditional probability,an approximate Bayesian calculation(ABC)method based on order statistics is introduced to estimate the parameters of SV model.Finally,the paper makes an empirical study based on the Shanghai composite index data,constructs the SV model by using the improved ABC method,ABC method and MCMC method,compares the models,and makes an empirical analysis on the fluctuation characteristics of China’s financial time series data.The main innovations of this paper are as follows: first,through Monte Carlo simulation,this paper compares and analyzes the SV model parameters obtained based on the improved ABC estimation method and the SV model parameters obtained based on ABC method and MCMC estimation method.The experimental results show that the improved ABC method can be better used for the parameter estimation and inference of SV model.Secondly,the paper makes an empirical analysis of China’s financial time series data by using the SV model.Through the DIC criterion,it is found that the fitting effect of the SV model estimated based on the improved ABC method is better,which shows that the approximate Bayesian theory can better realize the statistical inference of the random volatility model for risk management in the field of economy and finance,Predicting the fluctuation of asset price has certain application value.