Skip to main content

Table 3 Mean SC metric across all spring and fall testing corpora and training corpora sizes

From: Enhancing data pipelines for forecasting student performance: integrating feature selection with cross-validation

Testing corpus

Time frame

Feature selection technique

Number of training semesters

2

3

4

5

Spring

Pre

CAE

0.94

0.96

0.96

0.96

FSA

0.94

0.95

0.87

1.00

IG

0.86

0.92

0.93

0.81

RAE

0.30

0.19

0.24

0.16

Week 3

CAE

0.95

0.90

0.91

0.89

FSA

0.96

0.95

1.00

0.90

IG

0.82

0.89

0.88

0.74

RAE

0.17

0.17

0.18

0.13

Week 6

CAE

0.95

0.97

1.00

0.90

FSA

0.92

0.94

0.89

0.94

IG

0.91

0.85

0.85

0.79

RAE

0.20

0.19

0.19

0.16

Week 9

CAE

1.00

0.96

1.00

0.89

FSA

0.96

1.00

1.00

1.00

IG

0.93

0.88

0.88

0.81

RAE

0.19

0.17

0.25

0.24

Fall

Pre

CAE

0.93

1.00

1.00

 

FSA

0.92

1.00

1.00

IG

0.86

0.85

0.86

RAE

0.19

0.23

0.25

Week 3

CAE

0.93

1.00

0.87

FSA

0.97

1.00

1.00

IG

0.79

0.86

0.88

RAE

0.23

0.13

0.17

Week 6

CAE

1.00

1.00

0.96

FSA

0.97

0.93

1.00

IG

0.73

0.76

0.85

RAE

0.16

0.14

0.21

Week 9

CAE

0.93

1.00

0.90

FSA

0.95

1.00

1.00

IG

0.83

0.96

0.80

RAE

0.16

0.31

0.15

  1. Based on our prediction design, no fall testing corpus was available when five training semesters were used