Table 4.
Discriminant analysis using log-transformed LCA and GDCA as a model (GCDCA and GLCA were not significant and hence not entered into the model)
| Classification results | |||
|---|---|---|---|
| Group | Predicted group membership | ||
| Control | MCI | AD | |
| Control (N = 30) | 12 (40.0%) | 9 (30.0%) | 9 (30.0%) |
| MCI (N = 20) | 3 (15.0%) | 11 (55.0%) | 6 (30.0%) |
| AD (N = 30) | 10 (33.3%) | 5 (16.7%) | 15 (50.0%) |
| Goodness of prediction (in brackets: bias-correction by means of cross-classification) | |||
|---|---|---|---|
| Correct | One category offa | Completely misclassifiedb | |
| Control (N = 30) | 40.0% (40.0%) | 30.0% (30.0%) | 30.0% (30.0%) |
| MCI (N = 20) | 55.0% (50.0%) | 45.0% (50.0%) | – (–) |
| AD (N = 30) | 50.0% (46.7%) | 16.7% (20.0%) | 33.3% (33.3%) |
| Total (N = 80) | 47.5% (45.0%) | 28.8% (31.3%) | 23.7% (23.7%) |
Classification formula
Control f0 = 6.576* ln(GDCA) + 8.673* ln(LCA) – 35.996
MCI f1 = 5.636* ln(GDCA) + 10.451* ln(LCA) – 36.547
AD f2 = 6.528* ln(GDCA) + 10.356* ln(LCA) – 41.679
Rule Classify as control if f0 > f1 and f0 > f2, classify as MCI if f1 > f0 and f1 > f2, classify as AD if
f2 > f0 and f2 > f1
aActual group and predicted group in adjacent categories, e.g., actual group = healthy, predicted group = MCI
bActual group and predicted group are completely different, e.g., actual group = healthy, predicted group = AD