The Research On Prediction Of Gold Price Based On Parameter And Model Uncertainty

Gold,as a special commodity with financial attributes,is one of the more important spot and futures trading products in the commodity market.Gold is related to a country's financial security and an important part of a country's official reserves.The change of gold price is closely related to the financial situation at home and abroad,the value behavior of investors and producers.During the economic crisis in 2008,many financial assets showed abnormalities,except in the gold trading market.In the recent large fluctuations in China's stock market,the gold price did not show particularly obvious fluctuations.Based on large data,combined with time-varying parameters and dynamic forecasting method of the model,this paper explores the law of gold price change,and seeks the best gold price forecasting model to adapt to the changing economic environment to dynamically predict the trend of gold price.The analysis and Research on the factors affecting gold price will help investors better understand the gold market and provide help for investors and the state in decision-making.In this paper,dynamic model averaging(DMA)and dynamic model selection(DMS)methods are introduced to construct the dynamic prediction model of gold price.At the same time,the value of gold is decomposed into commodity attribute,currency attribute and hedging attribute variables.The CRB index,USDX index,exchange rate of US dollar against Japanese yen,Euro,British pound and VIX index are selected to forecast gold price.Measurement.It is found that commodity index and VIX index are positively correlated with gold price,while USDX index is negatively correlated with gold price.The fluctuation of gold price is affected by multiple factors.Two time-varying(TVP)models,Bayesian Model Average(BMA),Bayesian Model Selection(BMS),three Autoregressive(AR)models and Ordinary Least Square Linear Regression(OLS)models are selected as the comparative models to compare the prediction results.The empirical results show that DMA model is the best model for predicting gold price.Then,this paper analyses the advantag es of parameter and model uncertainty methods from the perspective of predictive force of predictive variables,expected number of predictive variables and time-varying forgetting factors,and combines with empirical analysis to analyze the influencing factors of gold volatility.Finally,based on the empirical results of this paper,the countermeasures and suggestions on gold allocation are put forward from the national level,the investor level and the technical level.