小麦赤霉病CBR预测模型参数的优化.PDFVIP

安徽农业大学学报, 2014, 41(1): 82-86 Journal of Anhui Agricultural University 网络出版时间：2013-12-17 16:15:49 [URL] /kcms/detail/34.1162.S1615.008.html 小麦赤霉病 CBR 预测模型参数的优化 1 1* 1 1 1 2 2 关 东 ，陈 莉 ，张沙沙 ，丁克坚 ，朱 诚 ，刘家成 ，郑朝阳 (1.安徽农业大学植物保护学院，合肥 230036 ；2.安徽省农业委员会，合肥 230061) 摘 要：基于类比推理人工智能方法 CBR 原理建立了以时间序列为轴线的小麦赤霉病滚动预测模型。为提高 模型预测的准确性，利用安徽省农作物病虫数据库，结合专家经验会商的结果，优化筛选该预测模型中的预测单元、 预测阶段、各预测阶段的权重、气象因子及其权重等各关键预测参数，并检验其预测的准确性。结果表明，运用优 化后的参数，小麦赤霉病 CBR 预测模型常年预测准确率可达 84.21%。故优化后的预测参数可用于小麦赤霉病 CBR 预测模型。 关键词：案例推理；预测因子；权重；优化 中图分类号：S435.121.45 文献标识码：A 文章编号：1672−352X (2014)01−0082−05 Parameter optimization for CBR model of wheat scab 1 1 1 1 1 2 2 GUAN Dong , CHEN Li , ZHANG Shasha , DING Kejian , ZHU Cheng , LIU Jiacheng , ZHENG Zhaoyang (1. School of Plant Protection, Anhui Agricultural University, Hefei 230036; 2. Agriculture Committee in Anhui Province, Hefei 230061) Abstract: A rolling forecasting model of wheat scab was set up using CBR theory with time series. To im- prove the accuracy and reliability of this model, impact factors and parameters, such as model unit, prediction stages, meteorological factors and its weight were optimized by using the database system of crops pests in Anhui province and the results of expert consultation. The accuracy of CBR model of wheat scab with optimized pa- rameters was tested, and the results showed that the precision rate was 84.21% in a normal year. The optimized factors and parameters could be applied to the forecasting model of wheat scab. Key words: case-based reasoning; impact factors; weight; optimization 赤霉病是我国小麦生产上重要的病害，其产生 的预测[8] 。李军等用地面气象因子、500 hPa 高空环 的毒素严重威胁人类和动物健康