高阶前馈神经网络在区域环境质量评价中的应用研究

Application of higher order feedforward neural networks in regional environmental quality assessment

目的 探索高阶前馈神经网络模型特性 ,将其应用于区域环境质量评价。方法 通过在多层前馈神经网络中增加高阶连接权建立高阶前馈神经网络模型 ,用此模型研究区域环境质量评价 ,并与传统BP网络的应用结果进行对比。结果 高阶前馈神经网络模型应用于区域环境质量评价时 ,其性能指标优于传统BP网络。

Objective To analyzehe characteristics of higher order feedforward neural networks model and demonstrate the application of the model in regional environmental quality assessment. Methods A new way to investigate regional environmental quality assessment by using higher order weights in multilayer feedforward neural networks was provided. The regional environmental quality assessed by using higher order feedforward neural networks model was compared with that by traditional BP model. Results The properties of higher order feedforward neural networks model were superior to those of the traditional BP model when applied in the assessment of regional environmental quality. Conclusion The fine features enable the higher order feedforward neural networks model to be of good prospects in regional environmental quality assessment.