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

Table 4

The classification accuracy (i.e., the proportion of all signal types correctly detected) using methods of FFPC and PCPEM with window size 𝑙=1280 for the epileptic seizure detection problem using EEGs from Freiburg dataset. The channels used for classification are randomly selected from those 6 channels.

Methods	2 Channels	3 Channels	4 Channels	5 Channels	6 Channels
FFPC
Patient 1	0.999±0.001	0.997±0.010	1.000±0.000	1.000±0.000	1.000±0.000
Patient 3	0.999±0.003	0.997±0.005	0.988±0.007	0.990±0.003	0.991±0.003
Average	𝟎.𝟗𝟗𝟗±𝟎.𝟎𝟎𝟐	𝟎.𝟗𝟗𝟕±𝟎.𝟎𝟎𝟕	𝟎.𝟗𝟗𝟒±𝟎.𝟎𝟎𝟑	𝟎.𝟗𝟗𝟓±𝟎.𝟎𝟎𝟏	𝟎.𝟗𝟗𝟓±𝟎.𝟎𝟎𝟏
PCPEM
Patient 1	0.970±0.031	0.957±0.032	0.958±0.027	0.974±0.015	0.984±0.002
Patient 3	0.950±0.022	0.895±0.043	0.899±0.023	0.902±0.018	0.892±0.005
Average	𝟎.𝟗𝟗𝟔±𝟎.𝟎𝟐𝟔	𝟎.𝟗𝟐𝟔±𝟎.𝟎𝟑𝟕	𝟎.𝟗𝟐𝟗±𝟎.𝟎𝟐𝟓	𝟎.𝟗𝟑𝟖±𝟎.𝟎𝟏𝟔	𝟎.𝟗𝟑𝟖±𝟎.𝟎𝟎𝟑