论文部分内容阅读
在滑动相关普查的基础上,分别建立以气象、海温和环流资料为预报因子的油菜菌核病病情指数预报子模式。对子模式采用平均、加权、回归和人工神经网络(ANN)方法进行综合集成。结果表明,集成模式提高了历史样本的拟合精度和独立样本试报的准确性,特别是人工神经网络集成模式的效果更令人满意。
Based on the slide-related census, sub-models for forecasting the disease index of rapeseed sclerotinia based on meteorological data, sea surface temperature and circulation data were established respectively. The sub-model using averaging, weight, regression and artificial neural network (ANN) method for integrated integration. The results show that the integrated model improves the accuracy of historical samples and the accuracy of independent samples. Especially the effect of artificial neural network integrated model is more satisfactory.