交叉熵神经网络及其在闽北大雨以上降水预报中的应用  

Application of BP neural network using cross-entropy to 96 hours forecast of heavy precipitation event in northern Fujian province

作　　者：吴木贵 江彩英[2] 张信华[2] 赖荣钦 

机构地区：[1]福建省建阳气象雷达站,建阳354200 [2]福建省南平市气象局,南平353000

出　　处：《南京信息工程大学学报（自然科学版）》2012年第3期220-225,共6页Journal of Nanjing University of Information Science & Technology（Natural Science Edition）

基　　金：福建省自然科学基金(2008J0241)

摘　　要：基于误差平方和最小化准则的BP神经网络(ANN-MSE)并不适合解决小概率天气事件的预报问题,引进一种改进的以交叉熵函数为目标函数的神经网络方法(ANN-CE),该法是一个三层反向传播神经网络,其输出层只用一个节点.利用2003—2008年的ECMWF预报场资料,把该法用于福建省南平市4—6月部分大雨或以上降水96h预报中,分别用原始因子和PCA降维后的主因子建立了ANN-CE预报模型和ANN-MSE预报模型,用这些模型对2009—2010年独立样本进行了试报.测试结果显示主因子预报模型TS评分比原始因子预报模型高且漏报次数少,其中,主因子ANN-CE预报模型的TS评分和漏报率分别是0.51和0.17,其性能是所有模型中最好且最为稳定的,是一种适合于小概率事件预报的方法.As a neural network based on MSSE,ANN-MSE is not an appropriate solution to the problem of predic- ting rare weather event. In this paper, an improved neural network method, ANN-CE is presented,which is a three layered back-propagation neural network with one output unit. The error function of ANN-CE is a cross entropy function. Then utilizing ECMWF forecast fields data, this method was applied to 96 hours foreeast of heavy preeipita- tion event in northern Fijian province. The ANN-CE model and the ANN-MSE model based on original faetor and principle component after PCA reducing dimensions were respeetively built. These models were applied to independ- ent samples in 2009--2010 ,and the test results are as following：TS grade for model based on prineipal component is higher than that of model based on original factors;miss-rate for the ANN-CE model is lower than that of the ANN-MSE model. All in all, ANN-CE model based on prineipal eomponent has best performance and stability, whose TS grade and miss-rate was respectively 0.51 and 0. 17 ,so it was suited for forecasting rare event.

关 键 词：BP神经网络 交叉熵 分类预报 小 概率事件 

分 类 号：TH71[机械工程—测试计量技术及仪器] TG803[机械工程—仪器科学与技术]