Skip to main content

Table 5 relative Root mean square error (rRMSE) value of tomato shoot mass (SM), fruit weight (FW), and number of fruits (FN) using random forest (RF), ridge regression (RI), and support vector machine (SVM) models with selected variables set from first-order statistics and all variables

From: Prediction of plant-level tomato biomass and yield using machine learning with unmanned aerial vehicle imagery

Model

From first-order statistics

Boruta

DALEX

GA

LASSO

RFE

SM [kg plant−1]

 RF

17.8

22.2

16.9

22.5

22.9

 RI

26.4

26.7

24.9

30.6

26.7

 SVM

17.6

18.9

16.7

21.4

21.8

FW [kg plant−1]

 RF

14.0

13.9

13.2

15.7

24.1

 RI

49.6

48.5

48.5

50.1

48.0

 SVM

14.3

14.5

14.6

18.7

15.9

FN [piece plant−1]

 RF

12.6

14.2

10.0

12.4

14.2

 RI

30.4

25.5

30.3

18.1

21.2

 SVM

13.1

14.2

13.5

13.0

13.6

Model

From first- and second-order statistics

Boruta

DALEX

Genetic

LASSO

RFE

SM [kg plant−1]

 RF

18.7

16.5

15.0

13.6

13.4

 RI

22.7

20.5

11.0

25.8

17.6

 SVM

21.4

18.7

11.2

21.6

20.5

FW [kg plant−1]

 RF

18.0

15.6

12.5

13.8

15.6

 RI

11.5

20.6

12.8

22.1

16.7

 SVM

14.7

15.4

14.6

15.9

14.5

FN [piece plant−1]

 RF

14.9

13.5

11.2

11.5

13.5

 RI

15.5

17.7

12.7

28.1

13.4

 SVM

12.8

10.9

8.8

10.6

14.1