Table 9 Ten-fold crossover results of the proposed bKSHHO-KELM method for predicting microseismic and blasting phenomena.

From: A prediction model for microseismic signals based on kernel extreme learning machine optimized by Harris Hawks algorithm

Fold

Selected

feature

subset size

Accuracy

Recall

Precision

F1 score

#1

4

98.319%

95.000%

100.000%

0.974

#2

14

94.118%

86.047%

97.368%

0.914

#3

8

97.479%

100.000%

92.105%

0.959

#4

9

97.479%

100.000%

92.105%

0.959

#5

12

94.958%

90.000%

94.737%

0.923

#6

6

94.068%

96.970%

84.211%

0.901

#7

5

96.639%

97.222%

92.105%

0.946

#8

14

95.798%

94.595%

92.105%

0.933

#9

10

94.958%

94.444%

89.474%

0.919

#10

9

92.437%

85.366%

92.105%

0.886

Avg.

/

95.625%

93.964%

92.632%

0.931

Std.

/

0.019

0.052

0.043

0.028