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

Table 2

The 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 problem using one hour long EEGs from Freiburg dataset.

Seizure detection	𝑙=512	𝑙=768	𝑙=1024	𝑙=1280
FFPC
Patient 1	0.998±0.001	1.000±0.000	1.000±0.000	1.000±0.000
Patient 3	0.994±0.017	1.000±0.000	0.996±0.009	1.000±0.000
PCPEM
Patient 1	0.986±0.022	0.963±0.043	0.983±0.023	0.980±0.026
Patient 3	0.987±0.022	0.992±0.008	0.980±0.020	0.987±0.019