一种用于机场气象预测的模糊神经网络模型

Fuzzy Neural Network Model for Weather Forecasting of Airport

针对民用机场多因素气象预测问题的复杂性,该文构建出一种基于粗糙集的模糊神经网络模型。采用粗糙集理论约简属性,挖掘潜在规则,在此基础上建立模糊神经网络模型,并根据规则的统计性质和离散化结果初始化网络参数,采用BP算法训练网络。实例验证,该模型在收敛速度与预测精度上优于传统的神经网络模型。

For a multifactor weather prediction problem, this paper constructs a new model of fuzzy neural network based on rough set. This model applies the rough set theory to reduce factors of weather and extract rules, the model of fuzzy neural network is built by these rules, and the initial parameters＇ value of fuzzy neural networks is decided by the statistical parameters of rules and the result of discretization and the networks is trained by BP methods. The application example showi that this model is superior to the traditional one on fitting precision and training rate.