The Research On The Forecasting Method Of International Gold Price

For a long time,gold has attracted much attention because of its functions such as value scale,circulation means,storage means,payment means and world currency.Moreover,due to the complex nature of gold,the price of gold is not only affected by the relationship between supply and demand,but also affected by the factors such as US dollar index,US dollar exchange rate,Dow Jones index and crude oil price.No matter the macro-control of the country or the investment behavior of the general investors,it is necessary to predict the gold price scientifically and carry out effective economic activities.Therefore,the in-depth study of the gold price will help us grasp the objective laws,adjust the investment strategies and control the financial risks,which has great theoretical and practical value.The innovation of this paper lies in two aspects.On the one hand,this paper uses the gray correlation analysis to analyze the factors that significantly affect the gold price,and add four establishes economic factors to the multi factor neural network model to improve prediction accuracy.On the other hand,the combination forecasting model based on genetic algorithm is established to improve the prediction ability of the model.The model has strong practical value in short-term prediction.As time series,gold price itself may contain linear and non-linear relations.Therefore,based on four factors that significantly affect gold price,this paper establishes five prediction models of gold price by analyzing the historical data of gold price,then this paper compares the prediction results and generalization performance of several models to identify the optimal prediction model.First of all,this paper establishes the traditional forecasting model ARIMA,determines the order of the model according to the criterion AIC,establishes the linear function of the current period and the lag period of gold price,and uses the software Eviews to train and forecast.Secondly,because of the formation process of gold price can be regarded as a grey system containing both known information and unknown information,this paper establishes the grey forecasting model GM(1,1).According to the processed original data,the parameters in the grey differential equation are solved to determine the model structure,and the software Matlab is used for training and forecasting.Then this paper digs the non-linear relationship in the gold price sequence to establish the neural network model.This paper selects the appropriate network structure and adds four factors of US dollar index,US dollar exchange rate,Dow Jones index and crude oil price to build a multi factor neural network model,then this paper uses Matlab software for training and forecasting,further reducing the error rate of the predicted value.Finally,this paper will combine ARIMA model,GM(1,1)model and multi factor neural network model to build a new combination forecasting model.The weights of three several single models are determined according to the genetic algorithm,and the prediction results of three several single models are weighted to get the final prediction results of the combined model.This paper's optimal forecasting model is the combination forecasting model by contrasting the empirical results of four single models and the combined model.The model combines the advantages of the single models and adds the most significant factors affecting the price of gold,which can accurately predict the trend of gold price.Based on this model,we can predict the gold price in the next five days according to the historical data of the first thirty days,and provide corresponding policy suggestions for individual investors,gold market producers,the government and the state.In the aspect of gold market investors,individual investors need to be cautious and rational investment.In the aspect of gold market producers,producers should adjust the production scale and quantity of gold products in time,improve the production profit and improve the development of gold market.In terms of government support,we should promote the production and R&D capacity of the gold industry,and learn from the advanced technological achievements and management experience of developed countries in the transformation of the gold industry.In terms of national gold reserves,the state should adjust the proportion of gold assets of China's international reserves in time,improve China's foreign exchange reserve structure,prevent economic risks and resist international economic turbulence.