Fig 1.
Damping oscillation curve.
Fig 2.
Gray wolf ranks.
Fig 3.
Trend about label (X) and feature (Y) (23 elements).
Fig 4.
Trends in label X and feature Y (top 10 elements).
Fig 5.
Trends in label X1 and feature Y1.
Fig 6.
a and b change trend chart.
Fig 7.
MKMD algorithm flow chart.
Fig 8.
Damping oscillation function (before adding random numbers).
Fig 9.
Damping oscillation function (after adding random numbers).
Fig 10.
MKMDIGWO algorithm framework.
Table 1.
Description of the data set used in the experiment.
Table 2.
Detailed parameter values of the mRMR+PSO algorithm.
Table 3.
Detailed parameter values of the mRMR+GA algorithm.
Table 4.
Detailed parameter values of the mRMR+BBA algorithm.
Table 5.
Detailed parameter values of the mRMR+CS algorithm.
Table 6.
The highest classification accuracy of each algorithm on each data set and its feature subset length.
Table 7.
Average classification accuracy and average feature subset length of each of the six algorithms on each data set.
Fig 11.
Classification accuracy of 100 iterations over 12 data sets.
Fig 12.
Feature subset length for 100 iterations over 12 data sets.
Fig 13.
Modulation of damping oscillation on the 11t data set.
Fig 14.
Modulation of damping oscillation on the BD data set.
Fig 15.
Modulation of damping oscillation on the BP data set.
Fig 16.
Modulation of damping oscillation on the DE data set.
Fig 17.
Modulation of damping oscillation on the GE data set.
Fig 18.
Modulation of damping oscillation on the LU data set.
Fig 19.
Modulation of damping oscillation on the SH data set.
Fig 20.
Modulation of damping oscillation on the EC data set.
Fig 21.
Modulation of damping oscillation on the PA data set.
Fig 22.
Modulation of damping oscillation on the PT data set.
Fig 23.
Modulation of damping oscillation on the SC data set.
Fig 24.
Modulation of damping oscillation on the WI data set.
Table 8.
The comparison based on Wilcoxon signed-rank test on SC data set.