2.1. Data Set
The data set used in the study consists of 165 organic compounds obtained from the literature (Duchowicz & Castro, 2002; Katritzky et al., 1998; Poling, Thomson, George, & Friend, 2008); this set is structurally very diverse, and it includes saturated unsaturated hydrocarbons, halogenated compounds, compounds containing hydroxyl, cyano, amino, ester, ether, carbonyl, and carboxyl functionalities. The data were split three times into the training, calibration, and validation sets (Toropov, Toropova, Benfenati, Leszczynska, & Leszczynski, 2009a, 2009b; Alla P. Toropova, Toropov, Benfenati, Leszczynska, & Leszczynski, 2015). The training set was used to build up a QSAR model, the calibration set was used as the “critic” of the developed QSAR model, and the test sets was used in assessed for the predictive power of the model (Toropova & Toropov, 2014). The Supplementary materials section (Table S1), contains SMILES, the corresponding experimental and calculated critical temperatures (Tc).