A novel SVM-GCM model for prediction of surface tension of some ionic liquids at different temperatures

نوع فایل:PDF

تعداد صفحات:

سال انتشار:1395

چکیده

In this study, a novel group contribution method (GCM) is developed to predict the surface tension of ionic liquids (ILs) using support vector machine (SVM) algorithms. A data set including 1225 experimental data for 133 ionic liquids collected from various references is used to develop and assess the predictive capability of the model. The data set covers a temperature range of 278.15 through 470.56 K and a surface tension range of 19.15 through 70.30 mN.m-1. It employs a total of 45 anion-cation groups plus temperature to predict the surface tension. To better distinct the effects of anion and cation on the surface tension of ionic liquids, 5 cation groups, 36 anion groups and 4 substituted functional groups attached to the cation group, were implemented. The results of this method show an average absolute relative deviation (AARD) of 3.48% from experimental data.

واژگان کلیدی