【摘要】 为改善TS模糊神经网络在台风灾情中的预测精度并减少运算时间,该文利用改进的TS模糊神经网络建立华南地区台风灾情预测模型.首先选取1981-2016年登陆华南地区的台风灾害历史数据,运用灰色关联度法选取承灾体因子、防灾减灾因子;其次利用主成分分析法对因子进行线性降维;然后引入模糊C-均值聚类(FCM)确定隶属度函数的中心;最后对华南地区台风灾情进行预测试验,并将结果与TS模糊神经网络、BP神经网络的试验结果进行比较,结果表明,改进的TS模糊神经网络预测精度最优,且训练预测用时较短.更多还原

【Abstract】 In order to improve the TS fuzzy neural network prediction accuracy in the typhoon disaster and reduce the computation time,this paper uses the improved TS fuzzy neural network to establish the typhoon disaster prediction model in South China.Firstly,historical data of typhoon disasters landed in South China from the year 1981 to 2016 are selected,and factors of disaster-relief body factor and disaster prevention and reduction factor are selected by using gray relational analysis method.Secondly,principal component analysis is used to linearly reduce the factors.Then,the fuzzy Cmeans clustering(FCM)is introduced to determine the center of membership function.Finally,the typhoon disaster prediction in southern China is tested and compared with TS fuzzy neural network and BP neural network.The experimental results show that the improved TS fuzzy neural network has the best prediction accuracy and the training prediction involves a short time.更多还原