高等学校化学学报

高等学校化学学报 ›› 1993, Vol. 14 ›› Issue (5): 618.

• 研究论文 • 上一篇    下一篇

用人工神经网络处理谷物成分分析

吉海彦, 严衍禄   

  1. 北京农业大学生物学院, 北京 100094
  • 收稿日期:1992-09-07 修回日期:1993-01-03 出版日期:1993-05-24 发布日期:1993-05-24
  • 通讯作者: 吉海彦

The Quantitative Analysis for Cereal s Components by Artificial Neural Networks

JI Hai-Yan, YAN Yan-Lu   

  1. Biology College, Beijing Agricultural University, Beijing, 100094
  • Received:1992-09-07 Revised:1993-01-03 Online:1993-05-24 Published:1993-05-24

PDF (PC)

被引次数

17

摘要： 本文用人工神经网络处理谷物的付立叶变换近红外漫反射光谱,对谷物中含量在10-1～10-3的蛋白质、脂肪和6种人体必需氨基酸定量分析数据进行了解析,分析结果与经典化学方法没有系统偏差,且优于逐步回归分析法的结果。

关键词: 人工神经网络, 定量分析, 谷物

Abstract: Cereal' s eight components (the concentration between10-1～10-3), that is, protein (Pro), fat(Fat), leucine(Leu), isoleucine(Ile), valine(Val), threonine(Thr), phenylalanine(Phe) and lysine(Lys) were quantitatively analyzed by Artificial Neural Networks.For the eight components the prediction correlation coefficients (R) are 0.969, 0.892, 0.897, 0.884, 0.970, 0,860, 0.854, 0.899, and the coefficients of variability are 2.66%, 5, 39%, 7.62%, 7.06%, 2.97%, 7, 83%, 13.38%, 2.68% respectively.These results are superior to those of SRA, and have no systematical errors with classical chemistry' s methods.

Key words: Artificial neural networks, Quantitative analysis, Cereal

TrendMD: