Table 5 Model accuracy and comparison.

From: Supervised machine learning models for predicting student mathematics performance in Somaliland primary examinations 2023

 

LR

 

NB

 

RF

 

DT

 

SVM

 

KNN

 

Predicted

Fail

Pass

Fail

Pass

Fail

Pass

Fail

Pass

Fail

Pass

Fail

Pass

Fail

2349

322

2656

15

2322

349

2285

386

2545

126

2598

73

Pass

541

978

40

1479

516

1003

532

987

1118

401

751

768

%

            

Accuracy

0.794

 

0.986

 

0.793

 

0.780

 

0.703

 

0.803

 

Sensitivity

0.812

 

0.985

 

0.818

 

0.811

 

0.694

 

0.775

 

Specificity

0.752

 

0.990

 

0.741

 

0.718

 

0.760

 

0.913

 

Pos pred value

0.879

 

0.994

 

0.869

 

0.855

 

0.952

 

0.972

 

Neg pred value

0.643

 

0.973

 

0.660

 

0.649

 

0.264

 

0.505

 

Prevalence

0.689

 

0.643

 

0.677

 

0.672

 

0.874

 

0.799

 

Detection rate

0.560

 

0.633

 

0.554

 

0.545

 

0.607

 

0.620

 

Detection prevalence

0.637

 

0.637

 

0.637

 

0.637

 

0.637

 

0.637

 

Balanced accuracy

0.782

 

0.987

 

0.780

 

0.765

 

0.727

 

0.844

 

AUC

0.856

 

0.999

 

0.856

 

0.806

 

0.745

 

0.868

 
Back to article page