Table 1 Prediction precision and accuracy of forage mass parameters (dry forage mass, dry leaf forage mass and dry green forage mass) of Marandu palisadegrass pastures using random forest and support vector regression models.

From: Using sentinel-2 satellite images and machine learning algorithms to predict tropical pasture forage mass, crude protein, and fiber content

Variables

Input features#

Linear regression

CCC

R2

RMSPE

Models

RF

SVR

RF

SVR

RF

SVR

Dry forage mass (g/m2)

Bd

0.20

(0.05)

0.26

(0.03)

109.68

(5.37)

105.40

(6.62)

0.37

(0.03)

0.44

(0.04)

Bd + Mt

0.35

(0.05)

0.35

(0.08)

98.77

(4.73)

100.0

(3.98)

0.52

(0.03)

0.54

(0.05)

VI

0.20

(0.06)

0.25

(0.08)

109.49

(5.44)

106.51

(6.30)

0.36

(0.04)

0.42

(0.05)

VI + Mt

0.34

(0.06)

0.37

(0.07)

99.22

(6.21)

97.04

(8.81)

0.52

(0.04)

0.52

(0.05)

Bd + VI

0.23

(0.06)

0.27

(0.07)

106.95

(6.47)

105.04

(5.73)

0.38

(0.04)

0.45

(0.04)

Bd + VI + Mt

0.34

(0.06)

0.38

(0.07)

99.81

(6.24)

96.57

(9.38)

0.50

(0.03)

0.52

(0.05)

Dry leaf forage mass (g/m2)

Bd

0.44

(0.08)

0.51

(0.07)

43.08

(2.19)

40.38

(2.31)

0.62

(0.06)

0.67

(0.04)

Bd + Mt

0.53

(0.08)

0.61

(0.07)

39.31

(2.44)

36.01

(3.75)

0.68

(0.07)

0.76

(0.04)

VI

0.45

(0.07)

0.63

(0.01)

42.74

(1.66)

35.19

(1.31)

0.62

(0.05)

0.77

(0.01)

VI + Mt

0.56

(0.05)

0.62

(0.02)

37.88

(1.17)

35.69

(1.88)

0.71

(0.04)

0.78

(0.02)

Bd + VI

0.46

(0.06)

0.49

(0.07)

42.11

(1.45)

41.33

(1.11)

0.64

(0.04)

0.67

(0.03)

Bd + VI + Mt

0.56

(0.04)

0.62

(0.03)

38.26

(1.48)

35.51

(1.91)

0.70

(0.03)

0.76

(0.02)

Dry green forage mass (g/m2)

Bd

0.36

(0.08)

0.51

(0.06)

83.67

(5.46)

73.97

(3.85)

0.52

(0.06)

0.66

(0.04)

Bd + Mt

0.49

(0.06)

0.58

(0.06)

74.41

(4.34)

67.67

(6.11)

0.65

(0.05)

0.73

(0.04)

VI

0.40

(0.08)

0.50

(0.06)

80.52

(4.35)

74.82

(3.60)

0.57

(0.06)

0.64

(0.05)

VI + Mt

0.52

(0.05)

0.64

(0.03)

72.52

(4.01)

63.48

(3.25)

0.67

(0.04)

0.76

(0.02)

Bd + VI

0.43

(0.08)

0.52

(0.08)

79.13

(6.38)

73.41

(4.69)

0.58

(0.06)

0.66

(0.05)

Bd + VI + Mt

0.52

(0.07)

0.63

(0.04)

72.55

(4.23)

63.54

(4.07)

0.66

(0.05)

0.76

(0.03)

  1. The number in parenthesis represents a standard error among fivefold cross-validation. Bold represents the best models.
  2. RMSPE root mean square prediction error, CCC concordance correlation coefficient, RF random forest, SVR support vector regression.
  3. #Bd: only data from spectral bands; see Table 5 for more information; Mt: meteorological data; maximum temperature (Tmax), minimum temperature (Tmin), average temperature (Tavg), relative humidity (RH), number of rainy days within a month (ND), rainfall; VI: only data from vegetation indices; see Table 6 for more information.
Back to article page