Why the study?
Heterogeneous EEG signals make feature extraction challenging, and prior emotion classification works using EEG data without removing heterogeneity have led to inaccurate classification.
Population
EEG data from two publicly accessible emotional datasets, AMIGOS and DREAMER
Comparison
EMD-based methodology with FFNN vs prior works without removing data heterogeneity
Key result
An EMD-based methodology using a Feed-Forward Neural Network for EEG data showed a 5-6% increment in accuracy, precision, and recall scores for emotion classification.
Authors
Loading...
May advance EEG emotion classification methods in research; leaves open validation for cardiovascular applications.
Effect estimate: 5-6% increment
An EMD-based methodology with a Feed-Forward Neural Network improves emotion classification accuracy from EEG data by 5-6%.
Gahlan et al. (2023) studied Emotion classification. EMD-based methodology with Feed-Forward Neural Network (FFNN) vs. Prior works was evaluated on Accuracy, precision, and recall scores for emotion classification (5-6% increment). An EMD-based methodology using a Feed-Forward Neural Network for EEG data showed a 5-6% increment in accuracy, precision, and recall scores for emotion classification.