Table 5 Pricing errors of models with and without alternative data.

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

NO.

Models

MSE

RMSE

MAE

Without

With

Δ

Without

With

Δ

Without

With

Δ

1

MLR

5.486

19.589

2.342

4.426

1.950

2.027

2

Lasso

5.688

4.602

2.385

2.145

1.983

1.763

3

DT

5.578

3.178

2.362

1.120

1.530

1.783

4

SVR

7.273

6.812

2.697

2.610

2.267

2.199

5

MLP

5.302

1.707

1.787

1.307

2.303

0.924

6

KNN

6.223

3.550

2.495

1.884

1.997

1.346

7

GBDT

2.815

1.863

1.678

1.365

1.292

1.013

8

LSTM

3.510

2.934

1.874

1.713

1.455

1.341

9

RF

2.419

1.168

1.555

1.081

1.143

0.766

10

LGBM

2.723

0.994

1.650

0.997

1.246

0.649

  1. The symbol Δ represents the difference in pricing errors between using alternative data and not using alternative data. If the difference is greater than zero, it is indicated with an , signifying that the alternative data has not reduced the pricing error. Conversely, if the difference is less than zero, it is indicated with a , demonstrating that the alternative data has effectively reduced the pricing error.
Back to article page