Skip to main content

Table 5 Performance of prediction models built with different VIs

From: Applying spectral fractal dimension index to predict the SPAD value of rice leaves under bacterial blight disease stress

VIs

Model

Training set

Test set

R2

RMSE

RE/%

R2

RMSE

RE/%

MSAVI

DT

0.8153

4.2358

9.3182

0.7916

4.7874

9.5617

PLSR

0.8019

5.2545

10.2113

0.7711

5.3593

10.3156

SVR

0.8553

4.0548

8.3935

0.8355

4.5187

9.3219

BPNN

0.8437

3.2254

8.7417

0.8322

3.3290

10.8533

MCARI

DT

0.7215

8.3541

14.0523

0.7006

9.2147

15.3319

PLSR

0.6853

7.4097

11.0561

0.6631

10.2416

12.1102

SVR

0.7783

6.8345

9.7764

0.7547

10.9724

9.1542

BPNN

0.7512

6.7714

10.2314

0.7431

9.2433

10.2011

MTCI

DT

0.5839

10.9318

20.2176

0.5581

18.5998

20.3154

PLSR

0.6337

13.4315

17.7154

0.6255

14.3392

18.6833

SVR

0.6239

8.3549

13.1171

0.6213

10.4582

14.6914

BPNN

0.6617

7.9018

12.1272

0.6571

9.8851

13.2387

RVI

DT

0.5311

12.2155

19.2513

0.4924

13.5217

19.7315

PLSR

0.5442

11.5125

18.2651

0.5351

12.3652

18.9113

SVR

0.5537

9.2254

13.7114

0.5463

10.3290

13.8151

BPNN

0.5329

8.7592

14.2615

0.5154

9.3651

14.3216

VARIred

DT

0.7419

9.8263

10.2344

0.7224

10.2355

12.9371

PLSR

0.7133

12.3652

14.2615

0.7062

13.6239

16.3117

SVR

0.7939

11.9217

10.0200

0.7819

12.9759

13.8592

BPNN

0.7785

8.8251

9.3154

0.7435

10.9472

9.8138

SFDI

DT

0.8413

4.5163

10.5127

0.8387

4.7184

10.6479

PLSR

0.8516

3.8715

9.8435

0.8479

4.5526

9.9316

SVR

0.8874

3.5124

7.7451

0.8752

3.7715

7.8614

BPNN

0.8759

3.3152

8.3218

0.8679

3.6780

8.6153

  1. The bold values highlight the best performance