摘要
以四川马尔康、甘肃天水两地1988 2000年小麦条锈病发生情况和期间的气象资料数据为基础,利用判别分析方法对小麦条锈病的发生程度进行预测,建立了判别函数,四川马尔康、甘肃天水数据资料回代检验错分率分别为0、0.153 8,交叉验证错分率分别为0.230 8、0.307 7。四川马尔康回代准确率为100%,交叉验证准确率81.82%;甘肃天水回代准确率为87.88%,交叉验证准确率为78.79%。可利用该方法作为小麦条锈病预测预报的参考,以指导小麦生产。
Wheat stripe rust, caused by Puccinia striiformis West. f. sp. tritici Eriks, is one of the pandemic diseases all over the world. In this article, forecast of wheat stripe rust was conducted based on the occurrence data of this disease and the climate data collected from Maerkang of Siehuan and Tianshui of Gansu during 1988 - 2000. The discrimination models were then built. The error-count estimates were 0 and 0. 153 8, and the posterior probability error-rate estimate was 0. 230 8 and 0. 307 7 for the data from Maerkang and Tianshui, respectively. The resubstitution accuracy and the cross-validation accuracy were 100 percent and 81.82 percent for the data from Maerkang. The resubstitution accuracy and the cross-validation accuracy were 87.88 percent and 78. 79 percent for the data from Tianshui. Therefore, the method of discrimination analysis could be a reference for the forecast of wheat stripe rust.
出处
《植物保护》
CAS
CSCD
北大核心
2006年第4期24-27,共4页
Plant Protection
基金
国家杰出青年科学基金--海外青年学者合作研究基金(30328018)
关键词
小麦条锈病
判别分析
预测
Wheat stripe rust
discrimination analysis
forecast
