Application Of Non-Parametric Method In The Analysis Of Gold Price Volatility And Correlation

At present,the process of financial globalization is accelerating,the types of financial institutions are continuously enriched,and financial tools are becoming more and more perfect.As an asset that can avoid risks,the price trend of gold is worthy of attention and research.In the financial market,gold price fluctuations are easily affected by external factors.Choosing an appropriate method to conduct in-depth research on gold price data is conducive to understanding the development status of the gold market.Nowadays,most literatures choose parameter methods when modeling and analyzing financial data,but the use of parameter methods for modeling needs to meet a series of assumptions,and the excellent properties of the parameters can only be reflected when the residual distribution of the regression model is known.The price of gold is related to a variety of factors and exhibits the characteristics of cluster fluctuations in time.Once the assumptions are not established or the model is set incorrectly,the estimation will be invalid.Therefore,the use of parametric modeling has shortcomings.Since the non-parametric method does not need to meet too many assumptions and does not pre-set the form of the model,it is based on the data generation process for modeling,which can accurately reflect the laws of economic operation.Therefore,the use of non-parametric method to model the data generation process based on gold price fluctuations has important theoretical significance and practical value.On the basis of existing research,this article uses non-parametric method to study the volatility and correlation of gold price.The theoretical part first summarizes non-parametric method,then introduces the characteristics of gold price data and common parametric volatility models,and finally studies the construction of non-parametric GARCH model.The empirical research is divided into two parts.The first part selects Au99.95 gold price return rate data,firstly establishes a parametric GARCH model and a non-parametric GARCH model,and conducts a comparative analysis to study the fluctuation law of gold price.The Bai-Perron test is used to detect structural mutation points of the return sequence,and perform segmented non-parametric estimation to further study the volatility.The second part selects the Au99.95 gold price and the exchange rate data of USD to RMB,uses the parametric method to obtain the influence path of the two,and uses the non-parametric method to deeply study the correlation between the gold price and the exchange rate.The innovation and empirical results of the article are embodied in the following two aspects: First,explore a non-parametric method to study the volatility and correlation of gold price,which provides a new research idea for the study of gold price.Second,combining the theory of non-parametric method with the characteristics of gold price volatility,it explains the value of non-parametric method to study gold price data from two perspectives of volatility and correlation.And the use of segmented non-parametric estimation to study the volatility of the gold price return sequence provides a realistic basis for the application of non-parametric method in the prediction of gold price volatility.The study found that when studying gold price data with complex volatility,non-parametric method,compared with parametric method,avoid the problem of model misconfiguration to a certain extent,reduce estimation errors,and have stronger interpretation capability.