Table 5 Modelling results for selected QSARINS models with Variable 3 along with their statistical validations.

From: Introduction of benzyloxy pharmacophore into aryl/heteroaryl chalcone motifs as a new class of monoamine oxidase B inhibitors

Statistical parameter

Model-9

Fitting

R2tr

0.9125

R2adj

0.8797

R2tr − R2adj

0.0328

LOF

0.2862

Kxx

0.0890

ΔK

0.2275

RMSEtr

0.2675

MAEtr

0.2015

RSStr

0.8586

CCCtr

0.9542

s

0.3276

F

27.8006

Internal validation

R2cv (Q2loo)

0.8347

R2 − R2cv

0.0777

RMSEcv

0.3676

MAEcv

0.2784

PRESScv

1.6212

CCCcv

0.9144

Q2LMO

0.7548

R2Yscr

0.2724

Q2Yscr

− 0.8777

External validation

RMSEex

0.4637

MAEex

0.4365

PRESSext

0.6450

R2ex

0.9157

Q2-F1

0.5618

Q2-F2

0.5429

Q2-F3

0.7370

CCCex

0.8672

Calc. external data regr. angle from diagonal

11.1830°

R2-ExPy (Predictions by LOO)

0.8372

R’o2

0.8223

k’

0.9984

r’2 m

0.7352

Ro2

0.8349

k

0.9979

r2m

0.7973

  1. *The statistical quality and strength of a GA-MLR based QSAR model was determined on the basis of: (a) internal validation based on leave-one-out (LOO) and leave-many-out (LMO) procedure (i.e. cross-validation (CV)); (b) using external validation; (c) Y-randomization (or Y-scrambling); and (d) fulfilling of respective threshold value for the statistical parameters: R2tr ≥ 0.6, Q2loo ≥ 0.5, Q2LMO ≥ 0.6, R2 > Q2, R2ex ≥ 0.6, RMSEtr < RMSEcv, ΔK ≥ 0.05, CCC ≥ 0.80, r2m ≥ 0.6, (1-r2/ro2) < 0.1, 0.9 ≤ k ≤ 1.1 or (1-r2/r’o2) < 0.1, 0.9 ≤ k’ ≤ 1.1,| ro2 − r’o2|< 0.3 with RMSE and MAE close to zero.
  2. Significant values are in bold.
Back to article page