基于BP神经网络的中、美贸易预测研究

根据中、美贸易的经济研究背景,分析中、美贸易重要影响指标,对其进行线性回归分析,发现中、美贸易额与各指标之间不是简单的线性相关,且各指标之间存在相互关系,所以不能用单一的线性回归模型进行预测。而BP神经网络具有非线性映射、自适应学习和良好的泛化能力等特征,运用BP神经网络模型对中、美贸易进行实证预测,大大提高了预测精度,取得了较好的效果。

Sino-American trade prediction research basing on BP nerve network

Some important influence indexes are analyzed according to the economical research background of Sino-American trade.Then they are carried on the regression analysis and discovered that they aren't the simple linear correlation between the Sino American volume of trade and various influence indexes.Therefore,it is not forecasted by the sole linear model.But the BP Neural Networks have the nonlinear mapping,the auto-adapted study and the good Generalization Ability etc.By applying the BP Neural Networks on the Sino-American trade,forecasting precision has greatly improved and the better effect has obtained.