Chem. J. Chinese Universities ›› 1994, Vol. 15 ›› Issue (12): 1775.

• Articles • Previous Articles     Next Articles

Neural Networks Used in Chromatography (Ⅰ)─prediction of GC Retention Parameters

LI Zhi-Liang1,2, O. Yaegashi2, LIANG Ben-Xi1, SHI Le-Ming3   

  1. 1. Hunan University, Changsha, 410082;
    2. Toyohashi University of Technology, Japan;
    3. Institute of Chemical Metallurgy, Academia Sinica
  • Received:1994-01-31 Revised:1994-07-05 Online:1994-12-24 Published:1994-12-24

Abstract: Neural networks algorithm was applied to chromatography. The relationship between GCretention parameters and induction effect I, hydrophobic index lgP, molar refractivity MRand molecular connection 1χ was studied by using the modified backpropagation (MBP) neural networks. The GCretention indices of alphatic amines, alcohols, acids and esters were estimated and predicted with relative error less than10%. The neural network method behaved as a good modelling technique for predicting GCretention parameters and might therefore be regarded as a powerful chemometric method.

Key words: Neural networks, Modified propagation algorithm, GC retention parameters

TrendMD: