Table 6 Pricing performances under different input features.

From: How to price a dataset: a deep learning framework for data monetization with alternative data

Methods

Features

80–20%

70–30%

90–10%

MSE

RMSE

MAE

MSE

RMSE

MAE

MSE

RMSE

MAE

LGBM

TF

2.7226

1.6500

1.2457

2.6122

1.6162

1.2317

2.5761

1.6050

1.2520

 

TF + Text (data title)

1.2715

1.1276

0.7684

1.5692

1.2527

0.8702

1.5022

1.2257

0.8198

 

TF + Text (target user)

2.6050

1.6140

1.1421

2.4328

1.1199

1.5597

2.7867

1.6694

1.2409

 

TF + Text (data function)

2.4041

1.5505

1.0854

2.2813

1.5104

1.0778

2.4970

1.5802

1.1401

 

TF + Text (data descriptions)

0.8016

0.8953

0.6248

0.9492

0.9743

0.6809

0.8116

0.9009

0.6173

 

TF + Text (total)

0.9941

0.9970

0.6489

1.1497

1.0722

0.7266

0.8526

0.9234

0.6389

MLP

TF

5.3025

1.7867

2.3027

5.4665

2.3381

1.8287

4.7131

2.1710

1.7092

 

TF + Text (data title)

2.6124

1.6163

1.1653

2.8345

1.6836

1.1803

3.2863

1.8128

1.2544

 

TF + Text (target user)

4.0210

2.0052

1.4659

3.8034

1.9502

1.4093

3.5362

1.8805

1.3539

 

TF + Text (data function)

4.4767

2.1158

1.6672

4.5859

2.1415

1.6115

3.2846

1.8124

1.3204

 

TF + Text (data descriptions)

1.0651

1.0320

0.6483

1.1049

1.0511

0.7034

0.9007

0.9491

0.6676

 

TF + Text (total)

1.7074

1.3067

0.9235

1.5713

1.2535

0.8629

0.9355

0.9672

0.7165

DT

TF

5.5779

2.3618

1.5295

5.2832

2.2985

1.5048

5.5555

2.3570

1.5112

 

TF + Text (data title)

4.8629

2.2052

1.4432

5.0400

2.2450

1.5140

4.2060

2.0509

1.2653

 

TF + Text (target user)

5.6352

2.3739

1.5087

4.5270

2.1277

1.3550

5.7507

2.3981

1.5182

 

TF + Text (data function)

4.1051

2.0261

1.2680

4.7251

2.1737

1.3668

5.2672

2.2950

1.5241

 

TF + Text (data descriptions)

2.9486

1.7171

1.0972

3.1107

1.7637

1.1673

3.1418

1.7725

1.1185

 

TF + Text (total)

3.1783

1.1199

1.7828

3.4990

1.8706

1.1850

3.1492

1.7746

1.0538

GBDT

TF

2.8153

1.6779

1.2922

2.8486

1.6878

1.2866

2.6176

1.6179

1.3153

 

TF + Text (data title)

1.8530

1.3612

1.0152

2.0159

1.4198

1.0273

1.8415

1.3570

0.9990

 

TF + Text (target user)

2.8905

1.7002

1.2572

2.7421

1.6559

1.2455

2.7182

1.6487

1.2887

 

TF + Text (data function)

2.8044

1.6746

1.2172

2.4539

1.5665

1.1528

2.3408

1.5300

1.1697

 

TF + Text (data descriptions)

1.1892

1.0905

0.8086

1.3966

1.1818

0.8818

1.2838

1.1331

0.8391

 

TF + Text (total)

1.8630

1.3649

1.0129

1.8482

1.3595

1.0199

1.5875

1.2600

0.9567

RF

TF

2.4185

1.5552

1.1429

2.4799

1.5748

1.1575

2.6446

1.6262

1.2125

 

TF + Text (data title)

1.7186

1.3110

0.9076

1.9694

1.4033

0.9913

1.7925

1.3388

0.9372

 

TF + Text (target user)

2.5307

1.5908

1.1290

2.3857

1.5446

1.0973

2.8026

1.6741

1.1928

 

TF + Text (data function)

2.3676

1.5387

1.0450

2.3092

1.5196

1.0264

2.5204

1.5876

1.0876

 

TF + Text (data descriptions)

1.1402

1.0678

0.7340

1.2443

1.1155

0.8120

1.0117

1.0058

0.6861

 

TF + Text (total)

1.1682

1.0808

0.7661

1.3727

1.1716

0.8350

1.0718

1.0353

0.7284

  1. TF means traditional features.
  2. The bold values indicate the best performance for each evaluation metric.
Back to article page