论文部分内容阅读
为了对生猪市场价格风险进行预警,根据我国2009年1月-2011年8月14个指标的32组样本数据,建立了广义回归神经网络(GRNN)预警模型,其中训练样本29组,测试样本3组.训练样本和测试样本的均方根误差、平均绝对误差(AAE)和相关系数都非常接近,说明建立的模型具有较强的泛化能力和鲁棒性,测试样本的AAE为0.0062,平均相对误差为2.3%,说明建立的GRNN模型具有很高的预测精度,可用于我国生猪市场价格风险预警研究和实际预测,并为政府有关部门指导生猪生产和进行市场调控提供决策依据.
In order to early warning the price risk in the pig market, a generalized regression neural network (GRNN) early-warning model was established based on 32 sample data of 14 indicators in China from January 2009 to August 2011, of which 29 training samples, 3 test samples Group.The root mean square error, average absolute error (AAE) and correlation coefficient of training samples and test samples are very close, which shows that the established model has a strong generalization ability and robustness, the AAE of the test sample is 0.0062, the average The relative error is 2.3%, which indicates that the established GRNN model has a high prediction accuracy and can be used in the research and actual prediction of the price risk warning in the pig market in our country. It also provides the decision-making basis for the government departments to guide the pig production and market regulation.