基于径向基概率神经网络的气象参数状态识别

Identifying the States of Meteorological Parameters Based on Radial Function Probabilistic Neural Network

局部气象参数实时数据是船舶航行、飞机起降所需要的非常重要的海洋气象参数。这些参数实时数据中的奇异数据对航行导航可能会导致危险后果。提出了一种基于径向基概率神经网络的气象重要参数状态识别方法用来识别奇异数据。将气象参数实时数据作为神经网络的输入,参数对应状态作为输出,通过对径向基概率神经网络模型训练,实现数据状态识别,将故障奇异数据进行有效识别。保证了气象观测系统输出数据的可靠性。实验结果表明,该方法可靠,且具有较好的泛化能力,能够实现气象参数实时数据状态的有效识别。

Local real time data of major meteorological parameters are very important and necessary for sailing of boats and ships, as well as taking off and landing of aircrafts. The odd data occurs in these real time data may bring unexpected and dangerous accident to navigation. A novel approach based on radial basis probabilistic neural network to identify these odd data is proposed. The real time data of meteorological pa- rameters are used as the input of the neural network, and the corresponding states are used as the output. Through radial function probabilistic neural network model training, state identification is implemented, and the reliability of the output data is guaranteed. The experimental result indicates that the method is reliable and is capable for real time data state identification and even for generalization.