高等学校化学学报

高等学校化学学报 ›› 1999, Vol. 20 ›› Issue (S1): 259.

• Chromatography and Separation Sciences • 上一篇    下一篇

Use of Artificial Neural Networks to Determine the Gas Chromatographic Retention Index data of Alkylbenzenes on Squalane (SQ)

YAN Ai-Xia, HU Zhi-De   

  1. Department of Chemistry, Lanzhou Universily, Lanzhou 730000, P. R. China
  • 出版日期:1999-12-31 发布日期:1999-12-31

Use of Artificial Neural Networks to Determine the Gas Chromatographic Retention Index data of Alkylbenzenes on Squalane (SQ)

YAN Ai-Xia, HU Zhi-De   

  1. Department of Chemistry, Lanzhou Universily, Lanzhou 730000, P. R. China
  • Online:1999-12-31 Published:1999-12-31

PDF (PC)

摘要：

Quantitative structure-property relationships (QSARs) quantify the connection between the structure and properties of molecules and allow the prediction of properties from structural parameters. Models of relationships between structure and retention index of aLkylbenzenes were constructed by means of a multilayer neural network using Extended Delta-Bar-Delta (EDBD) algorithms1'1. The 78 group data (taken from reference) belong to 33 alkylbenzenes under different temperatures on SQ. Considered,each of them has a same part of phenyl, each was uniquely presented by a set of numeric codes of 6 numbers depending on its substituents. Some examples were shown in Table 1. A set of six numbers and the temperature were used as input parameters to predict the retention indexes. The data were randomly divided into two sets:training set (60 members) and testing set (18 members) The structures of networks and the learning times were optimized. The best structure of network is 7-4-1 and the optimum learning times is about 750 epochs.

Abstract:

Quantitative structure-property relationships (QSARs) quantify the connection between the structure and properties of molecules and allow the prediction of properties from structural parameters. Models of relationships between structure and retention index of aLkylbenzenes were constructed by means of a multilayer neural network using Extended Delta-Bar-Delta (EDBD) algorithms1'1. The 78 group data (taken from reference) belong to 33 alkylbenzenes under different temperatures on SQ. Considered,each of them has a same part of phenyl, each was uniquely presented by a set of numeric codes of 6 numbers depending on its substituents. Some examples were shown in Table 1. A set of six numbers and the temperature were used as input parameters to predict the retention indexes. The data were randomly divided into two sets:training set (60 members) and testing set (18 members) The structures of networks and the learning times were optimized. The best structure of network is 7-4-1 and the optimum learning times is about 750 epochs.

TrendMD: