Skip to main content

Advertisement

Table 8 SVM multimodal classifier error rate for large set of ASSR ensemble averaged sweeps

From: 40-Hz ASSR fusion classification system for observing sleep patterns

Training subjects Test subject Training sweeps Test sweeps Error
A, B, C, D, E, F, G H 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 30%
A, B, C, D, E, F, H G 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 23%
A, B, C, D, E, G, H F 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 19%
A, B, C, D, F, G, H E 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 23%
A, B, C, E, F, G, H D 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 14%
A, B, D, E, F, G, H C 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 20%
A, C, D, E, F, G, H B 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 16%
B, C, D, E, F, G, H A 7,000(500 W0 + 500 N3) × 7 1,000(500 W0 + 500 N3) 24%
  1. Of different training and test subjects for channels C4-A1A2 and Fz-A1A2.