摘要
为给烟叶质量评价提供实用方法和理论依据,本研究基于模糊数学理论和神经网络机器学习理论,从大理红花大金元的主要化学成分和感官评吸结果确定烟叶品质的内在关系入手,通过BP算法建立了基于化学成分的神经网络模型。结果表明,模糊隶属函数结合BP神经网络的方法可以有效、可靠地运用于基于化学成分的烟叶质量感官评价辅助决策中,并可方便地应用于其它方面;还提供了BP神经网络的MATLAB实现程序。
A pattern recognition model for tobacco quality was established by using back-propagation （BP） algorithm in neural network,based on the relationship between tobacco quality with expert experience and the main chemical components in tobacco.The results showed that this algorithm could effectively and reliably be used in the tobacco sensory evaluation based on the main chemical components in tobacco and that it could also effectively be used in other fields.The designed Matlab programming was simply and useful by the result.
出处
《西南农业学报》
CSCD
北大核心
2012年第1期48-53,共6页
Southwest China Journal of Agricultural Sciences
基金
云南中烟工业公司资助项目(2008YL07)
红云红河烟草(集团)有限公司资助项目(YHH2010YL02
HYHH2010YL04)
关键词
神经网络
BP算法
烟叶质量
化学成分
感官评吸
Neural network
BP algorithm
Tobacco quality
Chemical components
Sensory quality
