摘要
通过探索原油的族组成与黏度的关系,比较了多元线性回归模型与BP网络模型对这种关系的关联能力及泛化能力.结果表明,BP网络模型对原油的族组成与黏度的预测远好于多元线性回归模型,其平均相对偏差仅为4.0%.
Through exploring the relationship between the 1 viscosity of crude oil and its group composition,it is compared that the capability of conrelation and generalization between the model of BP net and the model of multiple linear regression. The result shows the viscosity of crude oil can be predicted base on the former than the latter. The mean relative deviation ofthe former is only 4 %.
出处
《太原师范学院学报（自然科学版）》
2007年第2期12-14,共3页
Journal of Taiyuan Normal University:Natural Science Edition
关键词
原油
黏度
族组成
多元线性模型
BP网络
crude oil
viscosity,group-composition
multiple linear model
BP net
