Table 1 Comparison of original models for estimating the enthalpy of sublimation and models re-derived in the current study.

	Politzer et al.	Gharagheizi et al.	Bagheri et al.	Salahinejad et al.^a	Mathieu
Equation	5	6	7	10
Number of descriptors	2	5	3	4	35
Literature: ^b
Dataset sizes
(train.)	34	1079	1269	1042	814
(test)	5	269	317	260	486
R² (train.)	0.95	0.97	0.93	0.95	0.99
R² (test)	NA^c	0.97	0.93	0.95	0.99
Error^d	11.7	5.5	9.8	7.3	4.1
Re-derived here:^e
R²	0.82	0.97	0.96	0.89	0.70
Std. dev.^f	13.9	5.4	5.1	10.3	17.1
Largest deviation^g
Positive	127.7	15.8	14.5	34.7	132.1
Negative	−163.9	−31.5	−20.8	−90.4	−92.9

^aThis is a modified form of the original Salahinejad et al. model, with the W1 descriptor replaced by the Hy descriptor. ^bResults reported in the original analysis in the literature. ^cNot reported. ^dAverage error for the Politzer et al., root mean square error (RMSE) for Gharagheizi et al., Bagheri et al. and Mathieu, and standard error estimate (SEE) for Salahinejad et al. models in kJ mol⁻¹. ^eResults for the training set of 260 compounds from Salahinejad et al. re-derived here. ^fStandard deviation in kJ mol⁻¹. ^gDeviation from experiment in kJ mol⁻¹.

Quick links

Search