论文部分内容阅读
本文基于人工神经网络（ＢＰ） 方法，用毛细管色谱法预测汽油馏分的辛烷值，其预测最大绝对误差为０．２８ ，平均误差为０．１２２，比常用的线性回归数学模型法更能准确地预报辛烷值。
In this paper, based on the artificial neural network (BP) method, the octane number of the gasoline fraction is predicted by capillary chromatography. The predicted absolute error is 0.28 and the average error is 0.122, which is more accurate than the commonly used linear regression mathematical model Predict octane number.