Skip to main content

Table 5 REP for various models on the test set

From: Wheat physiology predictor: predicting physiological traits in wheat from hyperspectral reflectance measurements using deep learning

Model

LMA

Narea

SPAD

Nmass

Vcmax

Vcmax25

J

A

gs

Vcmax25/Narea

Mean

Multitask 1DCNN

7.611

8.446

7.430

8.764

18.723

14.004

10.417

14.485

32.590

15.897

13.837

Multitask 1DCNN, No Spectral Trim

7.026

7.463

6.887

7.610

18.907

14.294

10.252

13.586

31.792

15.243

13.306

Singletask 1DCNN

7.949

6.776

6.666

10.653

18.039

15.184

10.890

15.526

32.761

16.411

14.086

Singletask 1DCNN, No Spectral Trim

7.231

6.120

6.794

8.955

17.452

14.501

10.791

14.983

32.677

15.691

13.519

MLP

7.921

9.550

6.754

10.709

19.925

15.731

12.687

15.954

35.074

15.708

15.001

LSTM

11.337

14.058

9.089

11.222

25.840

17.848

14.338

17.269

37.151

22.766

18.092

PLS

7.076

7.650

7.264

9.179

19.181

17.477

10.639

16.069

34.937

14.350

14.382

XGBoost

9.425

7.849

7.630

10.314

24.261

17.011

13.144

16.320

37.131

19.289

16.237

Ensemble

6.767

6.543

6.655

8.188

16.843

14.549

9.763

13.902

31.965

14.301

12.948

  1. The results reported are the mean of three runs with different random seeds