Skip to main content

Table 2 Top five variables selected from first-order statistics and all variables by Boruta, DALEX, genetic algorithm (GA), least absolute shrinkage and selection operator (LASSO), and recursive feature elimination (RFE) for shoot mass (SM)

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

Rank

From first-order statistics

From first- and second-order statistics

Map

Variable

Date

Map

Variable

Date

Boruta

      

 1

Plant height

AVE

0626

Plant height

MOC-1

0712

 2

GNDVI

RANGE

0716

NDVI

SV

0712

 3

Plant height

MAX

0720

NDVI

DV

0712

 4

GNDVI

AVE

0712

Plant height

AVE

0702

 5

NDVI

SD

0712

GNDVI

DV

0724

DALEX

      

 1

Plant height

AVE

0626

NDVI

SV

0712

 2

Plant height

MAX

0702

Plant height

SE

0712

 3

Plant height

AVE

0702

NDVI

SE

0712

 4

GNDVI

MAX

0530

NDVI

DV

0712

 5

GNDVI

RANGE

0530

Plant height

Ent

0626

GA

      

 1

Plant height

RANGE

0618

Plant height

RANGE

0712

 2

Plant height

AVE

0626

NDVI

SV

0712

 3

GNDVI

RANGE

0716

NDVI

IDM

0702

 4

WDVI

MAX

0724

NDVI

DV

0712

 5

NDVI

SD

0712

GNDVI

MOC-2

0724

LASSO

      

 1

Plant height

AVE

0626

Plant height

AVE

0626

 2

GNDVI

RANGE

0716

GNDVI

DV

0724

 3

NDVI

MAX

0716

NDVI

SV

0716

 4

Plant height

MAX

0702

GNDVI

Con

0618

 5

Plant height

SKEW

0605

Plant height

MAX

0702

RFE

      

 1

Plant height

AVE

0626

NDVI

SV

0712

 2

NDVI

MAX

0716

Plant height

MOC-1

0712

 3

Plant height

MAX

0702

NDVI

DV

0712

 4

Plant height

AVE

0702

NDVI

MAX

0716

 5

NDVI

SD

0712

GNDVI

DV

0724