Table 7 Performance metrics for different modalities and classifiers with backward feature selection (Part 2).

From: Dynamically predicting comprehension difficulties through physiological data and intelligent wearables

Modality

PRV+EDA

Handcrafted features

Classifier

Accuracy

Precision

Recall

F1

Accuracy

Precision

Recall

F1

LR

\(57.28\%\pm 0.01\)

0.58

0.69

0.63

\(61.19\%\pm 0.21\)

0.62

0.68

0.65

LDA

\(58.44\%\pm 0.13\)

0.60

0.58

0.59

\(62.74\%\pm 0.84\)

0.56

0.67

0.61

GNB

\(57.78\%\pm 0.09\)

0.60

0.60

0.60

\(67.13\%\pm 0.83\)

0.61

0.66

0.63

Back to article page