人工神经网络方法预测气相色谱保留值

本文运用一典型的人工神经网络模型－“反向传播“模型的改进形式，研究了诱导效应指数Ｉ，摩尔折射度Ｒｏ，疏水亲脂参数ＩｇＰ，以及分子联通性指数与气象色谱保留行为的关系，实现了对色谱保留植的预测。神经网络预测模型的最大相对误差不超过８．７％。结果表明，该方法性能良好，可望成为色谱保留值预测的有效手段。

Prediction of Gas Chromatographic Retention Values by Artificial Neural Network

Based on an improved back-propagation model which is one of the typical artificial neural network, the relationship among induced effect I, molar refractivity R , hydrophobic parameter lgp,as well as molecular connection and chromatographic retention behaviour,has been studied. And retention data for the investigated solutes are predicted. The maximum relative error doesn't exceed 8.7%. The results show that the performance of the neural network approach is good, and therefore it might be referred as an effective assistant technique for prediction of gas chromatographic retention values.