Table 1. Machine Learning Models Developed in the Literature Using Molecular Descriptors as Inputs are Sorted by Year of Publicationa.
| refs | compounds family | property assessed | ML method |
|---|---|---|---|
| (48) | IL | density, molar liquid volume | MLR |
| (49) | solvents | polarizability | RBNN |
| (50) | IL | density | NN |
| (51) | IL | toxicological effect | MLP |
| (52) | organic solvents + IL | solvatochromic parameter | RBNN |
| (53) | IL | activity, enantio-selectivity | ANN + MLR |
| (54) | IL | viscosity | MLR + SVM |
| (55) | IL | heat capacity | MLR + ELM |
| (56) | IL | H2S solubility | QSPR + ELM |
| (57) | IL | ecotoxicity | MLR + MLP |
| (58) | IL | refractive index | ELM + MLR |
| (59) | IL | Henry’s law constant | MLR + SVM + ELM |
| (60) | IL | viscosity | ANN |
| (61) | DES | viscosity | MLR + ANN |
| (62) | cosmetic oils | viscosity | MLP |
| (63) | DES | viscosity, density | MLR |
| (64) | ester-alkane | mixing energy | ANN |
| (65) | DES | electrical conductivity | MLR |
| (66) | ES | density, viscosity | MLR |
| (67) | IL | viscosity, conductivity, density | SVR |
| (68) | ES | pH | MLR + ANN |
| (69) | DES | CO2 solubility | RF |
| (70) | chemicals | Abraham parameters, solvation free energy, solvation enthalpy | DNN |
| (71) | chemicals | molar mass, boiling temperature, vapor pressure, density, refractive index, aqueous solubility | DNN |
| (43) | F-refrigerants | vapor pressure | ANN |
| (72) | DES | pH | MLR + PLR + ANN |
| (73) | DES | surface tension | ANN |
| (74) | DES | electrical conductivity | ANN |
| (75) | IL + DES | infinite dilution activity coefficients | FM + DNN |
| (76) | polymers | glass transition, melting temperature | ANN |
| (77) | DES | eutectic composition, melting temperature | DT + MLR |
| (78) | IL | Henry’s law constants | SVM + RF + MLP |
| (79) | IL + DES | thermal conductivity | ANN |
| (80) | IL | surface tension, speed of sound | MLR + GBT |
| (81) | DES | CO2 solubility | ANN |
| (82) | DES | heat capacity | MNLR + ANN |
a
IL, ionic liquids; DES, deep eutectic solvents; MLR, multiple linear regression; RBNN, radial basis neural network; NN, neural network; MLP, multi-layer perceptron; ANN, artificial neural network; SVM, support vector machine; ELM, extreme learning machine; QSPR, quantitative structure–property relationship; RF, random forest; DNN, deep neural network; PLR, piecewise linear regression; FM, factorization machine; DT, decision trees; GBT, gradient boosting tree; MNLR, multiple nonlinear regression.