Steel Prices Index Prediction in China Based on BP Neural Network

Steel prices index in China are effected by upstream and downstream of steel supply chain, so it is necessary for steel prices index prediction to conduct correlation analysis between steel prices index and it’s influence factors. These influence factors are selected as input factors and steel prices index as output factor to establish a BPNN model, then the model is applied to predict with influence factors data ranging from October 2011 to October 2013 and output factor data ranging from November 2011 to November 2013. The training relative error is 0.32 %, and the prediction error is 6.8 %.The results prove BPNN has good predictive ability.