多种灾情预测模型的研究──2001年后上海市热带气旋灾情预测

Study on some disaster-forecasting models: prediction of disastrous losses due to tropical cyclones in Shanghai City after 2001

采用３种方法，即灰色模型ＧＭ、自回归－滑动平均模型ＡＲＭＡ和人工神经网络ＡＮＮ预测了上海市２００１年后（含２００１年）强热带气旋出现的趋势与规律。结果表明，２００１，２００６和２０１０年将出现中级以上灾情，而前两个年份灾情较重．通过数值计算结果比较了３种方法的应用特点，为进一步建立综合预测模型打下基础。

The trend and regularity of occurrence of strong tropical cyclones in Shanghai City after 2001 are predicted by three models (grey model, auto-regressive moving average model and artificial neural network model). The results show that there will be natural disaster conditions of greater than middle grade in 2001, 2006 and 2010, especially in first both years.The characteristics of three models in application are compared by numerical results to set up the base for establishing a comprehensive prediction model.