This work demonstrates that the transport properties of various organic substances in the gas phase, such as viscosity and thermal conductivity coefficients, can be estimated with acceptable accuracy by using 1D-QSPR models, which allow for the prediction of the studied property based solely on the chemical composition of the molecule. Here, we studied the transport properties (viscosity coefficient and thermal conductivity coefficient) of organic compounds for sufficiently representative database including approximately 5,000 carbon-, halogen-, oxygen-, nitrogen-, and sulfur-containing compounds. Using a simplex approach for modeling molecular structure and machine learning methods, such as multiple linear regression (MLR) and random forest (RF), adequate 1D QSPR models for the transport properties of individual substances in the gas phase were developed for the formed databases. Analysis of the influence of certain structural and physicochemical factors on the studied transport properties of organic compounds was carried out. Based on the developed 1D RF QSPR models, a computer expert system for predicting the viscosity coefficients and thermal conductivity of new substances was created.
Read full abstract7-days of FREE Audio papers, translation & more with Prime
7-days of FREE Prime access