|
| Table 5. Training versus test sets: validation results. |
SAR model |
Training |
Test |
||||||||
Intercept |
ASMmVQt |
lfDdOQg |
InMrLQg |
LsDMpQg |
ntr |
r2tr |
Ftr (p) |
nts |
r2ts |
Fts (p) |
-3.432 |
0.035 |
-0.341 |
0.087 |
-0.379 |
18 |
0.968 |
98 (1.4·10-9) |
12 |
0.980 |
74.92 (7.9·10-6) |
-3.300 |
0.034 |
-0.329 |
0.082 |
-0.352 |
19 |
0.975 |
139 (4.3·10-11) |
11 |
0.971 |
49.59 (9.9·10-5) |
-3.253 |
0.035 |
-0.321 |
0.075 |
-0.326 |
20 |
0.972 |
132 (1.7·10-11) |
10 |
0.974 |
43.28 (4.5·10-4) |
-3.175 |
0.034 |
-0.310 |
0.065 |
-0.328 |
21 |
0.986 |
275 (1.6·10-14) |
9 |
0.941 |
13.14 (1.4·10-2) |
-3.206 |
0.034 |
-0.320 |
0.078 |
-0.314 |
22 |
0.954 |
89 (3.7·10-11) |
8 |
0.994 |
114.7 (1.3·10-3) |
|
||||||||||
0.971 |
μ |
0.972 |
||||||||
0.011 |
StdDev |
0.019 |
||||||||
ntr =
number of compounds in training set; r2tr =
squared correlation coefficient of training set; Ftr =
Fisher parameter of training set; nts = number of compounds
of test set; r2ts = squared
correlation coefficient of test set; Fts = Fisher parameter
of test set; μ = arithmetic mean; StdDev = standard deviation;
p = type I error. |
|
|