Skip to main content
. 2021 Jul 13;15:659410. doi: 10.3389/fnhum.2021.659410

Table 5.

Ternary sentiment analysis results of the multi-modal model using the recurrent EEG decoding component.

Randomly initialized GloVe BERT
Model P R F1 (std) P R F1 (std) P R F1 (std)
Baseline 0.408 0.384 0.351 (0.07) 0.510 0.507 0.496 (0.06) 0.722 0.714 0.710 (0.05)
+ noise 0.373 0.399 0.344 (0.10) 0.531 0.519 0.504 (0.04) 0.711 0.706 0.700 (0.06)
+ ET 0.424 0.413 0.388 (0.06) 0.539 0.528 0.513 (0.04) 0.728 0.717 0.714 (0.05)
+ EEG full 0.391 0.387 0.353 (0.07) 0.505 0.505 0.488 (0.07) 0.724 0.715 0.711 (0.06)
+ EEG θ 0.397 0.409 0.360 (0.07) 0.516 0.510 0.498 (0.06) 0.715 0.708 0.704 (0.05)
+ EEG α 0.390 0.390 0.347 (0.08) 0.520 0.516 0.506 (0.05) 0.720 0.712 0.707 (0.05)
+ EEG β 0.350 0.370 0.302 (0.09) 0.523 0.519 0.509 (0.05) 0.732 0.720 0.717 (0.07)
+ EEG γ 0.409 0.397 0.359 (0.07) 0.517. 0.513 0.502 (0.04) 0.709 0.705 0.697 (0.06)
+θ+α+β+γ 0.401 0.400 0.368 (0.06) 0.522 0.516 0.505 (0.05) 0.722 0.717 0.713 (0.05)

We report precision (P), recall (R), F1-score and the standard deviation (std) between five runs. The best results per column are marked in bold.