Application Of BP And RBF Network Models In Gold Price Forecast

This thesis mainly uses the BP neural network and RBF neural network to forecast the gold price, and compares the performance of two kinds of models for predicting the gold price.First of all, we will analyze the gold price data in 2105 by power spectrum method and Lyapunov index. We draw a conclusion that gold price time series is nonlinear. Then we will reconstruct the space of the gold price and calculate the minimum embedding dimension. We will use the minimum embedding dimension as the number of neuron nodes. Finally, we will use BP and RBF neural network to simulate and forecast the data of gold price. By comparing the two kinds of forecast results, we find that both models can well predict the gold price, and we also find that RBF neural network model in the gold price forecast is more effective than BP neural network model.