基于神经网络的催化剂配方模型(英文)

CATALYST COMPOUNDING MODEL ON NEURAL NETWORK

本文应用人工神经网络中的BP网络与RBF网络,讨论脂肪醇催化剂配方的建模问题,并与传统的多元线性回归方法进行对比。结果表明,神经网络具有明显的优点,而RBF网络又比BF网络更加精确,且收敛速度要快1 000倍以上。

This paper deals with the model building on the compounding of catalyst fatty alcohol by BP network and RBF network and compares with the traditional multivariate linear regression method. The results indicate, neural network has clear merits and RBF network is more accurate than BP network, but the convergent speed is faster.