(a) Canonical variate analysis for discriminant functions. | |||||||
---|---|---|---|---|---|---|---|
Function | Eigenvalue | % of total dispersion | Canonical correlation | Wilk’s lambda | F | df | P |
1 | 15.52 | 83.4 | 0.969 | 0.013 | 41.9 | 20 | < 0.0001 |
2 | 2.98 | 16.0 | 0.865 | 0.225 | 16.0 | 12 | < 0.0001 |
3 | 0.09 | 0.5 | 0.299 | 0.898 | 1.8 | 6 | 0.11 |
4 | 0.01 | 0.1 | 0.114 | 0.986 | 0.6 | 2 | 0.52 |
(b) Classification matrix | |||||||
 |  |  | Predicted |  |  | % correctly classified | |
Actual | LL | LLR | LR | LRR | RR | Â | |
LL | 10 | 0 | 0 | 0 | 0 | 100.0 | |
LLR | 0 | 29 | 3 | 1 | 0 | 87.9 | |
LR | 0 | 3 | 35 | 3 | 0 | 85.4 | |
LRR | 0 | 0 | 2 | 10 | 0 | 83.3 | |
RR | 0 | 0 | 0 | 0 | 8 | 100.0 |