Prediction of auto-ignition temperatures of hydrocarbons based on support vector machine

被引：0

作者：

Pan, Yong ^{[1
]}

Jiang, Jun-Cheng ^{[1
]}

Cao, Hong-Yin ^{[1
]}

Wang, Rui ^{[1
]}

机构：

[1] Jiangsu Key Laboratory of Urban and Industrial Safety, College of Urban Construction and Safety Engineering, Nanjing University of Technology, Nanjing 210009, China

来源：

Shiyou Xuebao, Shiyou Jiagong/Acta Petrolei Sinica (Petroleum Processing Section) | 2009年 / 25卷 / 02期

关键词：

Linear regression - Hydrocarbons - Mean square error - Support vector machines - Molecular structure;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

The quantitative relationship existed between the auto-ignition temperature (AIT) and molecular structures of the hydrocarbon compounds were investigated based on the quantitative structure-property relationship (QSPR) study. The recently proposed support vector machine (SVM) method was applied to QSPR study of 90 hydrocarbon compounds, and the mathematics model for predicting the AIT of hydrocarbons based on the molecular structures was developed. Both internal and external validations were performed to validate the performance of the resulting models. The results showed that, the prediction results of SVM were in good agreement with the experimental values, with the average absolute error being 21.0C and the root mean square error being 27.21, which were superior to those obtained by multiple linear regression and neural network methods. This paper provides a new and effective method for predicting AIT of hydrocarbon compounds for engineering.

引用

页码：222 / 227