BP神经网络烟叶化学成分预测模型构建  

Establishment of Prediction Model for Tobacco Chemical Composition Based on BP Neural Network

作　　者：张明乾 刘立博 赵羡波 吴龙 张永安 林志华[2] 许山河 胡兴川 WU Long;ZHANG Yongan;LIN Zhihua;XU Shanhe;HU Xingchuan(Fujian Tobacco Industrial Co.,Ltd.,Xiamen,Fujian 361012;Longyan Branch,Fujian Tobacco Company,Longyan,Fujian 364000,China)

机构地区：[1]福建中烟工业有限责任公司,福建厦门361012 [2]福建省烟草公司龙岩市公司,福建龙岩364000

出　　处：《贵州农业科学》2020年第2期136-139,共4页Guizhou Agricultural Sciences

基　　金：福建中烟工业有限责任公司科技项目(FJZYKJJH2016004)

摘　　要：为烟叶化学成分预测和卷烟工业原料使用提供理论依据,应用回归分析法和BP神经网络分析2014-2017年龙岩永定土壤养分因子和烟叶化学成分的相关性,并构建预测模型。结果表明:采用回归模型预测龙岩永定烟叶化学成分和土壤养分指标无显著线性相关性;采用BP神经网络模型预测烟叶化学成分相关性和模型精准度较高。BP神经网络可为烟叶化学成分预测提供有效途径,具有较强的实用性。The regression analysis model and BP neural network model is respectively established according to the correlations between soil nutrients and tobacco chemical components from 2014 to 2017 to provide the theoretical basis for prediction of tobacco chemical composition and utilization of raw material in cigarette industry.Result:There are no significant linear correlations between soil nutrients and tobacco chemical components predicted by the regression analysis model.The prediction accuracy of correlations between soil nutrients and tobacco chemical components analyzed by the established BP neutral network model is higher.In conclusion,the BP neural network model with more practicality can provide an effective way for prediction of tobacco chemical composition.

关 键 词：烟叶 土壤养分 化学成分 BP神经网络 

分 类 号：S572[农业科学—烟草工业]