Skip to main content

Table 3 Modeling results based on spectral features combined with image features

From: Detection storage time of mild bruise’s yellow peaches using the combined hyperspectral imaging and machine learning method

Storage time (h)

Evaluation index

Classification model

RF

SVM (c = 1.0000, g = 0.1000)

XGBoost (eta = 1, max_depth = 1)

2

Number of misjudgments

0

1

0

Accuracy (%)

100.00

95.00

100.00

8

Number of misjudgments

1

5

1

Accuracy (%)

95.00

75.00

95.00

24

Number of misjudgments

4

11

3

Accuracy (%)

80.00

45.00

85.00

48

Number of misjudgments

4

12

4

Accuracy (%)

80.00

40.00

80.00

Total number of misjudgments

9

29

8

 

Overall accuracy (%)

88.75

63.75

90.00

 
  1. c” is the penalty coefficient, “g” is the radius of the kernel function, optimize “c” and “g” by Gridsearch
  2. “eta” is the learning rate, “max_depth” is the maximum depth of each tree