Skip to main content

Table 2 Internal validation of the five (5) QSAR models

From: Quantitative structure-activity relationship (QSAR) modelling study of some novel carboxamide series as new anti-tubercular agents

Parameters Model 1 Model 2 Model 3 Model 4 Model 5 Threshold
Friedman (LOF) 0.8669 0.9431 0.9408 0.9376 1.4731
R-squared (training set) 0.8563 0.8436 0.8440 0.8446 0.7234 > 0.6
Adjusted R-squared 0.8185 0.8025 0.8030 0.8037 0.6542 > 0.6
Cross validated (Q2cv) 0.7543 0.7315 0.7423 0.6620 0.5781 > 0.5
Standard error of estimation (SEE) 0.4336 0.4522 0.4517 0.4509 0.5671
Variance ratio (F value) 22.648 20.5100 20.5700 20.6538 10.4625
PRESS 3.5724 3.8866 3.8775 3.8638 6.4323 > 0.5
Average \( {\overline{R}}_m^2 \) (LOO-train) 0.6751 0.6501 0.6564 0.5721 0.4771 > 0.5
Delta \( {\overline{R}}_m^2 \) (LOO-train) 0.0898 0.0680 0.1184 0.0971 0.0755
Computed experimental error 0.1044 0.1044 0.1044 0.1044 0.1
RMSEP 1.0657 1.1101 1.1320 0.9995 0.3388
Y randomization (C\( {R}_{\mathrm{p}}^2\Big) \) 0.7609 0.7392 0.7468 0.7516 0.6296 > 0.5