Skip to main content
SpringerPlus

Table 7 Pairwise t tests on accuracy for different classification techniques

From: Comparing writing style feature-based classification methods for estimating user reputations in social media

 

F1

F1 + F2

F1 + F2 + F3

F1 + F2 + F3 + F4

t

p

t

p

t

p

t

p

(a) Segmenting type s = like

 Base versus Base

  C4.5 < NN

0.1168

0.9074

−0.2642

0.7926

−0.2391

0.8119

2.0894

0.0411

  C4.5 < SVM

0.4121

0.6818

−0.1558

0.8768

−0.1733

0.8630

1.9855

0.0519

  C4.5 < NB

0.4411

0.6608

−1.6644

0.1015

−1.7821

0.0800

0.7397

0.4625

  NN < SVM

0.2948

0.7692

0.1085

0.9140

0.0661

0.9475

−0.0980

0.9222

  NN < NB

0.3237

0.7473

−1.3992

0.1671

−1.5415

0.1287

−1.3530

0.1813

  SVM < NB

0.0287

0.9772

−1.5085

0.1369

−1.6103

0.1128

−1.2512

0.2159

 Ensemble versus Ensemble

  RS-C4.5 < RS-NN

0.0985

0.9219

0.0305

0.9758

0.0407

0.9677

0.8816

0.3816

  RS-C4.5 < RS-SVM

0.2781

0.7819

−0.0970

0.9231

−0.0561

0.9554

3.3050

0.0017

  RS-C4.5 < RS-NB

0.3815

0.7043

−3.1268

0.0028

−3.1920

0.0023

1.5449

0.1278

  RS-NN < RS-SVM

0.1791

0.8585

−0.1273

0.8992

−0.0967

0.9233

2.4291

0.0183

  RS-NN < RS-NB

0.2820

0.7790

−3.1492

0.0026

−3.2248

0.0021

0.6608

0.5113

  RS-SVM < RS-NB

0.1027

0.9186

−3.0297

0.0037

−3.1341

0.0027

−1.7781

0.0807

 Base versus Ensemble

  C4.5 < RS-C4.5

1.5089

0.1368

3.4529

0.0011

3.4001

0.0012

4.9580

0.0000

  NN < RS-NN

1.4840

0.1432

3.7333

0.0004

3.6663

0.0005

3.7754

0.0004

  SVM < RS-SVM

1.3702

0.1759

3.5103

0.0009

3.5164

0.0009

6.2156

0.0000

  NB < RS-NB

1.4470

0.1533

2.0191

0.0482

2.0188

0.0482

5.7430

0.0000

(b) Segmenting type s = dislike

 Base versus Base

  C4.5 < NN

−0.0517

0.9590

−0.3484

0.7288

−0.3112

0.7568

1.3922

0.1692

  C4.5 < SVM

−1.5922

0.1168

−0.3674

0.7146

−0.3674

0.7146

2.2502

0.0283

  C4.5 < NB

−2.1459

0.0361

−13.8448

0.0000

−13.8448

0.0000

0.1839

0.8548

  NN < SVM

−1.5282

0.1319

−0.0185

0.9853

−0.0555

0.9560

0.8677

0.3892

  NN < NB

−2.0790

0.0421

−13.5724

0.0000

−13.5858

0.0000

−1.2081

0.2319

  SVM < NB

−0.5914

0.5565

−13.5946

0.0000

−13.5946

0.0000

−2.0673

0.0432

 Ensemble versus Ensemble

  RS-C4.5 < RS-NN

−0.4462

0.6571

0.8472

0.4004

1.3534

0.1812

2.1927

0.0324

  RS-C4.5 < RS-SVM

−0.7100

0.4806

0.8186

0.4164

0.8827

0.3810

2.5443

0.0137

  RS-C4.5 < RS-NB

−12.1209

0.0000

−7.4664

0.0000

−7.4322

0.0000

0.5340

0.5954

  RS-NN < RS-SVM

−0.2651

0.7919

−0.0315

0.9750

−0.4770

0.6352

0.3537

0.7249

  RS-NN < RS-NB

−11.8060

0.0000

−8.2529

0.0000

−8.6805

0.0000

−1.6677

0.1008

  RS-SVM < RS-NB

−11.5687

0.0000

−8.2615

0.0000

−8.2942

0.0000

−2.0211

0.0479

 Base versus Ensemble

  C4.5 < RS-C4.5

1.0205

0.3118

2.2381

0.0291

2.2048

0.0315

4.0928

0.0001

  NN < RS-NN

0.6221

0.5363

3.4212

0.0012

3.8486

0.0003

4.8828

0.0000

  SVM < RS-SVM

1.8912

0.0636

3.4260

0.0011

3.4584

0.0010

4.3731

0.0001

  NB < RS-NB

−9.2234

0.0000

9.7104

0.0000

9.7104

0.0000

4.4482

0.0000

(c) Segmenting type s = sum

 Base versus Base

  C4.5 < NN

−7.3255

0.0000

−4.9494

0.0000

−4.8977

0.0000

−4.0895

0.0001

  C4.5 < SVM

−16.3176

0.0000

−9.1771

0.0000

−9.0680

0.0000

0.2961

0.7682

  C4.5 < NB

−9.5597

0.0000

−8.9258

0.0000

−8.8171

0.0000

−0.7126

0.4793

  NN < SVM

−7.9144

0.0000

−4.7452

0.0000

−4.8774

0.0000

4.3562

0.0001

  NN < NB

−0.7139

0.4790

−3.8878

0.0004

−4.0311

0.0002

2.6493

0.0105

  SVM < NB

10.5096

0.0000

2.0088

0.0506

2.0637

0.0450

−0.9480

0.3476

 Ensemble versus Ensemble

  RS-C4.5 < RS-NN

−0.3712

0.7119

−12.3659

0.0000

−12.6411

0.0000

−24.3113

0.0000

  RS-C4.5 < RS-SVM

−3.6321

0.0006

−0.7347

0.4656

0.3079

0.7593

8.8124

0.0000

  RS-C4.5 < RS-NB

−8.3171

0.0000

−5.1562

0.0000

−4.8929

0.0000

10.1076

0.0000

  RS-NN < RS-SVM

−3.2133

0.0022

10.4931

0.0000

11.6334

0.0000

27.1599

0.0000

  RS-NN < RS-NB

−7.8203

0.0000

11.0239

0.0000

10.8831

0.0000

32.8773

0.0000

  RS-SVM < RS-NB

−5.2629

0.0000

−3.5703

0.0010

−4.4613

0.0001

−2.0300

0.0485

 Base versus Ensemble

  C4.5 < RS-C4.5

−2.1194

0.0384

3.9081

0.0003

3.4528

0.0012

16.1968

0.0000

  NN < RS-NN

4.5425

0.0000

−1.8326

0.0720

−1.6728

0.0998

−8.4251

0.0000

  SVM < RS-SVM

11.4120

0.0000

13.3717

0.0000

14.9017

0.0000

20.2885

0.0000

  NB < RS-NB

−2.2939

0.0272

20.6640

0.0000

20.6640

0.0000

22.3312

0.0000

(d) Segmenting type s = portfolio

 Base versus Base

  C4.5 < NN

−0.3357

0.7398

−1.2074

0.2385

−3.2526

0.0028

1.1173

0.2723

  C4.5 < SVM

0.0746

0.9411

2.4769

0.0197

2.5035

0.0185

3.4966

0.0016

  C4.5 < NB

−2.7119

0.0115

−3.7995

0.0009

−3.8622

0.0008

6.9459

0.0000

  NN < SVM

0.5537

0.5836

5.0627

0.0000

6.9119

0.0000

2.4446

0.0206

  NN < NB

−3.2000

0.0031

−3.8374

0.0006

−0.1194

0.9059

6.1578

0.0000

  SVM < NB

−3.7482

0.0007

−8.8516

0.0000

−8.9767

0.0000

4.2191

0.0002

 Ensemble versus Ensemble

  RS-C4.5 < RS-NN

−5.9980

0.0000

−6.8700

0.0000

−6.9289

0.0000

9.1409

0.0000

  RS-C4.5 < RS-SVM

0.8009

0.4304

4.4411

0.0001

−0.4066

0.6870

17.5249

0.0000

  RS-C4.5 < RS-NB

2.7431

0.0110

4.0034

0.0004

2.8443

0.0080

12.4036

0.0000

  RS-NN < RS-SVM

8.1948

0.0000

10.5347

0.0000

6.4963

0.0000

11.5459

0.0000

  RS-NN < RS-NB

10.3702

0.0000

10.5375

0.0000

10.7623

0.0000

4.2133

0.0002

  RS-SVM < RS-NB

2.6804

0.0115

−0.8192

0.4189

3.2946

0.0026

−7.8686

0.0000

 Base versus Ensemble

  C4.5 < RS-C4.5

1.8607

0.0720

9.2073

0.0000

9.4357

0.0000

9.5237

0.0000

  NN < RS-NN

−4.7197

0.0001

5.8578

0.0000

8.0650

0.0000

21.0127

0.0000

  SVM < RS-SVM

4.0544

0.0003

12.7166

0.0000

8.4720

0.0000

34.1855

0.0000

  NB < RS-NB

10.7037

0.0000

25.1731

0.0000

25.7502

0.0000

26.3625

0.0000

  1. The results are t and p values of the t tests for classification technique comparison, and the results more than 5 % of significance level are highlighted in italics
Back to article page