基于Elman神经网络的气温预测研究

[目的]建立预测气温动态变化的Elman神经网络模型。[方法]参考温度的固有特性,应用1951～2010年的重庆温度数据,采用Elman人工神经网络模型对温度进行预测估计。[结果]实证仿真结果表明,模型的相对误差较小,能够对未来气温变化进行较好的模拟。[结论]该模型预测结果可以用于指导农业生产,并可以进一步应用到天气衍生品定价等领域。

Research on Temperature Forecasts Based on Elman Neural Network

[Objective] The aim was to establish Elman neural network model to predict the dynamic changes of temperature.[Method] Considering the inherent nature of temperature,and by dint of the temperature in Chongqing during 1951-2010,the Elman artificial neural network model was applied to predict the temperature.[Result] This simulation result suggested that the relative error was small and can have a good simulation to the future temperature changes.[Conclusion] The prediction result can guide agricultural production and further apply to the field of pricing the weather derivative products.