摘要
本文提出了以非线性回归拟合发展趋势、正弦函数逼近周期变动和马尔可夫链刻划随机扰动的集成预测模型，并应用于山东省农业自然灾害成灾面积的变动规律模拟，得到了较好的预测效果。理论和实践表明，集成预测模型优于传统的单模型预测，为预测具有复杂机制的自然灾害演变提供了一种新方法。
In this paper the integrated model of natural disaster prediction is proposed,in which,nonlinear model is used to simulate the trend,sine curve to fit the cyclical change and Markov chain to depict random disturbance.The integrated model is applied to predict the dynamic change of destroyed area by agricultural natural disaster in Shandong Province and the excellent results are obtained.It is proved by theory and practice that the integrated model of natural disaster prediction is better than the traditional single prediction models for predicting complex disaster.
出处
《自然灾害学报》
CSCD
1997年第3期16-21,共6页
Journal of Natural Disasters
