Research Article
Multiclass Informative Instance Transfer Learning Framework for Motor Imagery-Based Brain-Computer Interface
Table 1
Number of samples to reach baseline for different methods.
| Sub | Base | MIITAL | Reduction by MIITAL | Optimized | Reduction by optimized |
| A01 | 288 | 180 | 37.5% | 40 | 86% | A02 | 288 | 180 | 37.5% | 90 | 68.75% | A03 | 288 | 140 | 51.4% | 40 | 86% | A04 | 288 | | | | | A05 | 288 | | | 120 | 58% | A06 | 288 | 170 | 41% | 40 | 86% | A07 | 288 | | | 100 | 65.3% | A08 | 288 | 60 | 79% | 40 | 86% | A09 | 288 | 150 | 48% | 70 | 75.7% |
| Mean | 288 | ā | 49% | ā | 76.56% |
|
|
: baseline cannot be reached. |