Meteorological frontal surface automatic identification method based on multiple regression

The invention provides a meteorological frontal surface automatic identification method based on multiple regression. The method comprises: determining a constant temperature zone; removing the constant temperature zone; after a constant temperature zone is removed from the meteorological element data set, calculating each diagnosis quantity, and generating a training sample set; by utilizing a multiple regression neural network (MRNN) model and the training sample set, training a multiple regression equation coefficient (the weight of each diagnosis quantity during frontal surface recognition); and substituting the meteorological element data set and the multiple regression model coefficient obtained by training into a multiple regression model, and automatically identifying the atmospheric frontal surface probability of each grid point on line. 本发明提供了一种基于多元回归的气象锋面自动识别方法，包括：确定常定温带；去除常定温带；气象要素数据集去除常定温带后计算各诊断量，并生成训练样本集；利用多元回归神经网络(Multiple Regression Neural Network，MRNN)模型和训练样本集训练多元回归方程系数(各诊断量识别锋面时的权重)；将气象要素数据集