An Empirical Analysis Of London Gold Price-related Factors

With the diversification of financial products in the financial market, as well as financial derivatives diversification, there will be more and more investment in financial products, but in view of the characteristics of financial products, profit associated with risk, showed that the importance of accurate prediction. Many scholars began to predict the price of gold, the price of gold sequence is influenced by many factors, and has a strong random fluctuations and spikes, with a single time series models such as R, MA, ARIMA,but tend to be less satisfactory, the selected wavelet neural network to predict the price of gold. One hand, the use of wavelet denoising data smooth large noise, so get rid of the interference, the other hand, the neural network has a strong nonlinear fitting, can be mapped arbitrarily complex nonlinear relationship, so the neural network plus wavelet denoising based on the volatility of the price of gold will be able to be better prediction. So as to provide a reference for investors.The main contents are as follows:1. After studying the book about analysis of the elements that might affect the price of gold, and using PERSON correlation coefficient analysis,finding the dollar index, crude oil, silver, the Dow Jones index, the Shanghai Composite Index five identified, and to determine their relevance coefficient.2. Denoise the selected London gold price data, then put the dollar index, crude oil, silver, the Dow Jones index, the Shanghai index data into the go the established neural network model to denify the predictive value and real value for comparison.3. Then calculate the undenoising model and a denoising model’s RMSE MAXAE, MAE, MAXAPE MAPE with d wavelet neural network, to testify the wavelet neural network fit for the London gold price forcasting.