BP 神经网络模型在环境气象灾害预测中的应用研究--以湖南省为例

依据环境气象数据与自然灾害统计数据，建立BP神经网络模型，对湖南主要气象灾害（洪灾、旱灾、冰冻灾）及受灾经济损失进行实例预测，将在MATLAB7软件中的仿真结果与传统的多元线性回归模型分析结果进行比较和误差分析。结果表明，BP神经网络模型在洪灾、旱灾受灾率方面的预测效果和精度优于多元回归模型，而由于冰灾训练样本不足及经济损失与输入因子的线性相关程度高，在冰灾与受灾经济损失率方面稍逊于多元回归模型。

Application of BP Neural Network in Environmental Meteorological Disasters Prediction in Hunan Province

Using the latest predicting method of artificial neural network , this study builds up the topology of BP Neural Net-work based on data from Hunan Statistical Yearbook and other relevant sources.The program was simulated with Mat lab.The pa-per also establishes the Multiple Elements Linear Regression Model and ran in SPSS to predict the disasters .A comparison were made to analyze on the results of the two different models .The results show that BP neural network outperforms Multiple Elements Linear Regression Model in prediction for flood and drought disasters.But the BP neural network underperformed Multiple Ele-ments Linear Regression Model in predicting snow disaster and economy losses due to the lack of sufficient data inputs .