Table 17 Comparative analysis of the proposed DE against other basic and metaheuristic algorithms on the LFW dataset.

From: Reinforcement learning-driven feature selection enhanced by an evolutionary approach tuning for criminal suspect identification

Model

Accuracy

F-measure

G-means

AUC

Random search

69.214 ± 0.049

71.389 ± 0.052

72.119 ± 0.081

0.665 ± 0.061

Grid search

70.767 ± 0.045

72.396 ± 0.066

73.158 ± 0.038

0.672 ± 0.001

Hyperband

76.507 ± 0.004

78.865 ± 0.009

79.618 ± 0.008

0.681 ± 0.084

BO

81.400 ± 0.010

83.935 ± 0.010

84.716 ± 0.013

0.792 ± 0.020

SSA

74.546 ± 0.091

80.403 ± 0.074

81.019 ± 0.038

0.722 ± 0.052

HMS

75.851 ± 0.003

81.628 ± 0.006

82.178 ± 0.095

0.731 ± 0.089

COA

77.299 ± 0.007

83.101 ± 0.074

83.628 ± 0.021

0.746 ± 0.075

FA

78.137 ± 0.095

84.180 ± 0.003

84.690 ± 0.096

0.763 ± 0.098

BA

79.575 ± 0.008

85.533 ± 0.026

86.056 ± 0.089

0.780 ± 0.041

ABC

81.243 ± 0.056

86.489 ± 0.090

87.046 ± 0.014

0.789 ± 0.002

DE

82.324 ± 0.083

88.023 ± 0.054

88.542 ± 0.080

0.800 ± 0.039

Proposed DE

89.141 ± 0.022

91.152 ± 0.042

92.718 ± 0.074

0.845 ± 0.045