Skip to main content
. Author manuscript; available in PMC: 2013 May 25.
Published in final edited form as: J Chem Inf Model. 2012 Apr 27;52(5):1222–1237. doi: 10.1021/ci300092s

Table 3.

Ecom50 Molconn PLS and MLR modeling results (protonated and neutral structures)

Model na r2b sc Fd q2e spressf inputsg
Protonated Structures
PLS 52 0.931 0.46 84.3 0.910 0.53 7
MLR 52 0.862 0.65 34.3 0.812 0.76 8
Neutral Structures
PLS 52 0.905 0.54 59.9 0.855 0.67 7
MLR 52 0.848 0.68 30.7 0.762 0.90 8
a

Number of compounds.

b

Correlation coefficient for training set calculations.

c

Standard error (RMSE) for training set calculations.

d

Fisher F-test for training set calculations.

e

Correlation coefficient for leave-one-out cross validation.

f

Standard error for leave-one-out cross validation predications.

g

Number of descriptor inputs.