汉中地区小麦条锈病的BP神经网络预测 |
投稿时间:2000-03-22 点此下载全文 |
引用本文:胡小平,杨之为,李振岐,邓志勇,柯长华.汉中地区小麦条锈病的BP神经网络预测[J].西北农业学报,2000,9(3):28~31 |
DOI: |
摘要点击次数: 58 |
全文下载次数: 62 |
|
基金项目:国家自然科学基金资助项目(39770486) |
|
中文摘要:人工神经网络具有对非线性系统预测的重要特性,其跟踪性能好,适用面广,收敛速度快,容错能力强.本研究利用陕西汉中地区1974~1997年的病情、菌量、品种和气象资料,采用逐步回归法选择了影响汉中小麦条锈病流行的主要因子,即春季菌量、秋季菌量、感病品种面积比例、4月份降雨量和4月份平均温度,并将其作为BP神经网络预测模型的输入,用1974~1993年的资料进行网络训练,对1994~1997年小麦条锈病的流行程度作短期预测,结果高度吻合. |
中文关键词:小麦条锈病 BP神经网络 预测 汉中地区 |
|
Prediction of Wheat Stripe Rust in Hanzhong Area by BP Neural Network |
|
|
Abstract:A non-linear predictive algorithms based on back propagation neural network was used to predict the epidemics of wheat stripe rust in Hanzhong Area.This method has convergent speed, perfect trace performance, and high tolerant errors as well, especially, it is suitable for complex non-linear prediction.Factors affecting the wheat stripe rust epidemics in Hanzhong area were studied on the basis of the data of disease, pathogen, varieties and weather by regression method.The results showed that the key factors affecting the epidemics of wheat stripe rust are the density of pathogen in Spring, the density of pathogen in Autumn, the ratio of susceptible varieties, precipitation in April and daily average temperature in April.BP neural network model has been set up through selecting factors based on the data of 1974~1993.Short-term prediction of 1994~1997 has been carried out respectively.Results showed that the accurate ratio of prediction is high.The prediction model has bright application future. |
keywords:Wheat stripe rust (Puccinia striiformis West.) BP neural network Factors analysis Prediction Hanzhong area |
查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
|
|