应用人工神经网络对气相色谱 Kovats保留指数进行多点预测研究。基于人工神经网络的典型模型 (BP网络 ) ,采用文献中的数据 ,建立了反映溶质的 Kovats指数与溶质的溶剂化参数以及其与固定相的相参数之间关系的数学模型 ,并采用多元线性回归方法和人工神经网络对溶质的 Kovats指数进行了预测研究 Applying Artificial Neural Network to Predict Kovats Retention Index of Gas Chromatography by Multi-Point Prediction. Based on the data in the literature, a typical model based on artificial neural network (BP network) was established. The mathematical model reflecting the relation between the Kovats index of solute and the solvating parameter of solute and the phase parameter of solute with phase stationary phase was established. Regression method and artificial neural network to predict the solute Kovats index