论文部分内容阅读
我国是一个农业大国,农业受灾严重影响到国家和社会经济的可持续发展。文中以1970-2014年的农业总体、干旱和洪涝受灾面积为原始序列,提出了一种集成NARX神经网络和灰色系统的灾害预测模型。鉴于原始序列随机波动性较大的特点,对其乘以一个序列算子,得到一个相对平滑的新序列;采用灰色系统对新序列进行预测;用NARX网络对序列算子进行预测;将灰色预测值除以NARX网络预测值得到最终的预测结果。通过对农业总体、干旱和洪涝受灾面积进行实证研究,证实了这种预测方法能够有效地提高随机波动性较大序列的预测精度。
Our country is a big agricultural country, and the agricultural disaster has seriously affected the sustainable development of the country and society and economy. In this paper, a series of disaster prediction models integrating NARX neural network and gray system are proposed based on the original sequence of agricultural total, drought and flood affected area from 1970 to 2014. In view of the large variance of the original sequence, the new sequence is obtained by multiplying it by a sequence operator, using a gray system to predict the new sequence, using the NARX network to predict the sequence operator, Divide the value by the NARX network forecast to get the final forecast. Empirical research on the affected areas of agricultural total, drought and floods confirms that this prediction method can effectively improve the prediction accuracy of large random sequences.