波谱学杂志 ›› 1997, Vol. 14 ›› Issue (5): 403-412.

• 研究论文 • 上一篇    下一篇

神经网络用于核磁共振碳谱的研究:烷烃的化学位移和CSS与分子距边矢量λ

胡芳1, 刘树深1,2, 余般梅3, 彭升阳3, 莫立宇3, 曹晨忠1,4, 村松由起5, 李志良1,5   

  1. 1 湖南大学化学化工学院药物研究所, 长沙 410082;
    2 桂林工学院应用化学系, 桂林 541004;
    3 国防科技大学应用物理系, 长沙4 10073;
    4 湘潭师范学院化学系, 湘潭 411100;
    5 早稻田大学化学与生化系, 东京178
  • 收稿日期:1997-05-09 修回日期:1997-06-19 出版日期:1997-10-05 发布日期:2018-01-22
  • 作者简介:胡芳,女,24岁,硕士,助教
  • 基金资助:
    国家教委与自然科学基金和机械部跨世纪人才专项基金及日本文部省与科学振兴会资助项目

NEURAL NETWORK APPLIED TO THE STUDY OF CHEMICAL SHIFT OF 13C NMR SPECTROSCOPY IN ALKANES WITH A NOVEL MOLECULAR DISTANCEEDGE VECTOR (λ)

Hu Fang1, Liu Shushen1,2, Yu Banmei3, Peng Shengyang3, Mo Liyu3, Cao Chenzhong1,4, Muramatsu Y5, Li Zhiliang1,5   

  1. 1 Institute of Chemistry and Chemical Engineering, ICP, Hunan University, Changsha 410082;
    2 Department of Applied Chemistry, Gnilin Institute of Technology, Guilin 541004;
    3 Department of Applied Physics, Changsha Institute of Technology, Changsha 410073;
    4 Department of Chemistry, Xiangtan Teacher's College, Xiangtan 411100;
    5 Department of Chemistry and Biochemistry, Waseda University, Tokyo 178, Japan
  • Received:1997-05-09 Revised:1997-06-19 Online:1997-10-05 Published:2018-01-22

摘要: 系统研究了核磁共振碳谱及其化学位移规律性.提出了一种新的分子图论参数:分子距离-边数矢量(λ矢量),并发现了它与烷烃的13C NMR有良好的相关性.进一步用神经网络改进反传算法(BPNN)进行准确估计与定量预测,结果良好.

关键词: 核磁共振碳谱, 化学位移和CSS, 分子距边矢量λ, 烷烃, 神经网络, 改进反传算法

Abstract: Systematic studies were made on the 13C NMR and its regularity of chemical shift sum (CSS).In this paper, a set of novel molecular graph theoretical parameters, called the distance-edge vector (λ), was developed and found to be excellently correlated to 13C NMR CSS of alkanes. The modified backpropagation (MBP) naural networks were also applied to estimate and predict, the ressults satisfactoryly.

Key words: 13C NMR, CSS, Novel distance-edge vector (λ), Alkanes, Neural networks, BMP