| 基于烟草中致香成分的人工神经网络分类模型 |
| |
| 引用本文： | 陈清,刘巍,钟科军.基于烟草中致香成分的人工神经网络分类模型[J].湖南大学学报(自然科学版),2006,33(2):103-105. |
| |
| 作者姓名： | 陈清 刘巍 钟科军 |
| |
| 作者单位： | 1. 湖南大学,材料科学与工程学院,湖南,长沙,410082
2. 常德卷烟厂,湖南,常德,415000 |
| |
| 基金项目： | 教育部科学技术研究重点项目(03123)，湖南省烟草专卖局项目(03-007) |
| |
| 摘 要： | 从烟叶化学成分和结合评吸结果确定的烟叶品质的关系入手,用遗传BP神经网络建立了基于烟草中致香物质含量的烟草模式识别模型.分析结果表明,使用该方法能全局优化基于烟草致香成分的神经网络结构,较之梯度下降法的神经网络的训练效率更高,得到的识别结果也更加准确.它可以作为一种新的判别烟草香味特征的识别手段.
|
| 关 键 词： | 烟草 香味物质 遗传-神经网络 识别模型 |
| 文章编号： | 1000-2472（2006）02-0103-03 |
| 收稿时间： | 08 29 2005 12:00AM |
| 修稿时间： | 2005-08-29 |
A Neural Network Recognition Model Based on Aroma Components in Tobacco |
| |
| CHEN Qing,LIU Wei,ZHONG Ke-jun.A Neural Network Recognition Model Based on Aroma Components in Tobacco[J].Journal of Hunan University(Naturnal Science),2006,33(2):103-105. |
| |
| Authors: | CHEN Qing LIU Wei ZHONG Ke-jun |
| |
| Institution: | 1. College of Material Science and Engineering , Hunan Univ, Changsha, Hunan 410082, China; 2. Changde Cigarette Factory, Changde, Hunan 415000, China |
| |
| Abstract: | A pattern recognition model for tobacco quality was established by using back-propagation(BP) neural network combined with genetic algorithm,based on the relationship between tobacco quality with expert experience and aroma chemical component in tobacco.Experiment results showed that the neural network structure of tobacco aroma components was global optimized,and the neural network trained by genetic algorithm was more efficient and correct than by gradient descent method.It can be utilized as a new recognition method for identifying the aromatic characteristics of tobacco. |
| |
| Keywords: | tobacco aroma components genetic-neural network recognition model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《湖南大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《湖南大学学报(自然科学版)》下载免费的PDF全文 |
