Table 7 Comparison between proposed framework and baseline SEA study.

From: Beyond peak accuracy: a stability-centric framework for reliable multimodal student engagement assessment

Model

Proposed model

Baseline Model

Accuracy

Macro F1

Precision

Recall

Accuracy

Precision

Recall

Macro F1

Ensemble

0.9760

0.9690

0.9764

0.9756

–

–

–

–

Encoder

0.9510

0.9330

0.9544

0.9512

0.9300

0.9300

0.9300

0.8800

Inception

0.9510

0.9190

0.9543

0.9512

–

–

–

–

Snapshot

0.9270

0.8640

0.9377

0.9268

–

–

–

–

Transformer

0.9510

0.9260

0.9561

0.9512

–

–

–

–

FCN

0.9510

0.9460

0.9621

0.9512

0.9500

0.9500

0.9500

0.9100

TimeCNN

0.9510

0.9410

0.9512

0.9512

0.8500

0.7500

0.8500

0.5800

MCNN

0.9760

0.9690

0.9764

0.9756

0.8800

0.9000

0.8800

0.8500

Back to article page