Skip to main content

Table 5 AUC results for Train_Test

From: The effects of class rarity on the evaluation of supervised healthcare fraud detection models

Learner

Ratio

200

400

1000

All

(a) Part B

GBT

[Full]

0.77604

0.78207

0.79080

0.78636

[1:99]

0.78453

0.79440

0.80158

0.79523

[10:90]

0.78575

0.79954

0.81324

0.81566

[25:75]

0.75562

0.78570

0.80824

0.81476

[35:65]

0.74548

0.78412

0.80523

0.81057

[50:50]

0.72767

0.76659

0.79002

0.80626

LR

[Full]

0.80406

0.81686

0.82063

0.82133

[1:99]

0.80803

0.82062

0.82441

0.82542

[10:90]

0.80501

0.81933

0.82601

0.82901

[25:75]

0.79251

0.82101

0.82347

0.82675

[35:65]

0.77845

0.81197

0.82049

0.82768

[50:50]

0.76491

0.80030

0.81863

0.82109

RF

[Full]

0.70754

0.73765

0.75130

0.75725

[1:99]

0.73952

0.75813

0.76869

0.77081

[10:90]

0.76287

0.77653

0.78856

0.78527

[25:75]

0.75999

0.77324

0.78079

0.78683

[35:65]

0.75139

0.76720

0.77863

0.78505

[50:50]

0.74320

0.76689

0.77646

0.78136

Learner

Ratio

100

200

400

All

(b) Part D

GBT

[Full]

0.69437

0.70885

0.74099

0.74969

[1:99]

0.69356

0.71879

0.74821

0.76157

[10:90]

0.68142

0.70006

0.75261

0.78212

[25:75]

0.65184

0.69628

0.73418

0.77174

[35:65]

0.64126

0.67560

0.71313

0.77054

[50:50]

0.62193

0.65445

0.70517

0.74277

LR

[Full]

0.74339

0.76832

0.78592

0.79566

[1:99]

0.73766

0.76943

0.78623

0.79714

[10:90]

0.72541

0.76436

0.78491

0.79858

[25:75]

0.71602

0.75240

0.77726

0.79431

[35:65]

0.69825

0.74023

0.77341

0.79031

[50:50]

0.68920

0.72922

0.75804

0.78866

RF

[Full]

0.60202

0.62445

0.64317

0.69302

[1:99]

0.66243

0.68387

0.69370

0.73303

[10:90]

0.70282

0.72050

0.73803

0.77433

[25:75]

0.69181

0.71398

0.74121

0.76349

[35:65]

0.68118

0.70447

0.73372

0.75418

[50:50]

0.66406

0.68312

0.71714

0.75602

(c) DMEPOS

GBT

[Full]

0.72688

0.75808

0.78221

0.78281

[1:99]

0.73083

0.75805

0.78426

0.79202

[10:90]

0.71749

0.74800

0.77617

0.79683

[25:75]

0.67911

0.73027

0.76314

0.78660

[35:65]

0.66527

0.69776

0.75773

0.77678

[50:50]

0.65155

0.66424

0.74161

0.75944

LR

[Full]

0.75220

0.76024

0.77622

0.78088

[1:99]

0.74545

0.75646

0.77403

0.77819

[10:90]

0.73002

0.75079

0.76687

0.77482

[25:75]

0.70978

0.73345

0.75993

0.76963

[35:65]

0.68578

0.72187

0.75741

0.76715

[50:50]

0.67933

0.70508

0.74394

0.75723

RF

[Full]

0.65576

0.71649

0.75220

0.77105

[1:99]

0.67756

0.72877

0.76250

0.78803

[10:90]

0.70214

0.73717

0.77153

0.78861

[25:75]

0.71244

0.74737

0.76301

0.78914

[35:65]

0.70956

0.72159

0.76735

0.78076

[50:50]

0.69299

0.73423

0.75302

0.77333

Learner

Ratio

100

200

All

 

(d) Combined

GBT

[Full]

0.76056

0.78431

0.83654

 

[1:99]

0.75698

0.79823

0.84513

 

[10:90]

0.74038

0.79609

0.84929

 

[25:75]

0.73296

0.78145

0.83126

 

[35:65]

0.69390

0.77744

0.82757

 

[50:50]

0.70015

0.75962

0.81149

 

LR

[Full]

0.81514

0.85430

0.86888

 

[1:99]

0.80496

0.84899

0.86829

 

[10:90]

0.76965

0.82737

0.86157

 

[25:75]

0.73072

0.80287

0.84583

 

[35:65]

0.71753

0.77810

0.83743

 

[50:50]

0.69273

0.74712

0.81778

 

RF

[Full]

0.62150

0.72501

0.80122

 

[1:99]

0.71805

0.78308

0.82193

 

[10:90]

0.74251

0.78836

0.82896

 

[25:75]

0.74442

0.77432

0.81791

 

[35:65]

0.73639

0.77206

0.81273

 

[50:50]

0.72878

0.76666

0.81375

Â