Dynamic Principal Component Analysis with Nonoverlapping Moving Window and Its Applications to Epileptic EEG Classification

Table 1

The average classification accuracy (i.e., the proportion of all signal types correctly detected) using methods of FFPC and PCPEM with different values of 𝑙 for the epileptic seizure detection and diagnosis problems using the EEGs dataset from the University of Bonn.

Seizure detection	𝑙=64	𝑙=128	𝑙=256	𝑙=512
FFPC	0.973±0.003	0.982±0.003	0.991±0.001	0.994±0.001
PCPEM	0.999±0.001	1.000±0.000	1.000±0.000	1.000±0.000
Epilepsy diagnosis	𝑙=64	𝑙=128	𝑙=256	𝑙=512
FFPC	0.699±0.003	0.779±0.036	0.959±0.007	0.995±0.009
PCPEM	0.732±0.006	0.806±0.003	0.978±0.030	1.000±0.000