GCM 및 SVR 기반 회귀모델을 이용한 방향족 화합물 물성치 예측에 관한 연구

Abstract

In order to simulate a process model in the field of chemical engineering, it is very important to identify the physical properties of new material as well as existing materials. However, it is really hard to measure physical properties throughout experiments due to potential risk and cost.

To solve this difficulty, this study aims to develop a property prediction model based on group contribution method(GCM) for aromatic chemical compounds including benzene rings. The benzene rings of aromatic materials have a great impact on their physical properties. To establish the prediction model, 42 important functional groups determining physical properties are selected and their number on 147 aromatic chemical compounds is counted to prepare a data set. A support vector regression (SVR) is employed to make a prediction model since this data set is really sparse. To verify the efficacy of this study, the results of this study are compared with ones of previous researches. Even if the data set of previous researches are different each other, it is certain that the comparisons show that the performance of this study is better than others. Also, there is few research to predict physical properties of aromatic compounds. This study can provide an effective way to estimate physical properties of unknown chemical compounds and contribute to reducing the efforts of experiments for measuring physical properties.