Table 2.

A summary of the best-performing models.

Focus	Best Model	F2 Score
EEG-based Engagement	Random Forests (with Gini criteria, number of estimators = 1000)	0.86
Facial expression-based Engagement	Convolutional Neural Network (See Figure 5 for the architecture)	0.82
EEG-based Instantaneous Attention	Random Forests (with Gini criteria, number of estimators = 100)	0.82
Facial expression-based Instantaneous Attention	Random Forests (with Gini criteria, number of estimators = 1000)	0.81
EEG-based Focused Attention	Multilayer Perceptron (a single layer with 50 nodes and a logistic activation)	0.63
Facial expression-based Focused Attention	Multilayer Perceptron	0.53
EEG-based Planning	Extra Trees (with Gini criteria and number of estimators = 400)	0.68
Facial expression-based Planning	Extra Trees	0.78
EEG-based Shifting	Extra Trees (with Gini criteria and number of estimators = 100)	0.68
Facial expression-based Shifting	Extra Trees	0.81