Research Article

Trust in Intrusion Detection Systems: An Investigation of Performance Analysis for Machine Learning and Deep Learning Models

Table 12

The performance of deep learning for WSN dataset using LSTM and GRU with all features.

ModelEvaluation metricCross-validation performanceTesting performance
NormalGrayholeBlackholeSchedulingFloodingNormalGrayholeBlackholeSchedulingFlooding

LSTM with one layerTN99.9999.9810099.9910098.7199.9099.3592.9099.92
FP0.010.0200.0101.290.100.657.100.08
FN00.050.263.1923.337.056.6443.770.6310.66
Accuracy99.9999.9810099.9810098.5599.8497.6798.7799.73
Precision10099.9210096.769566.4788.9577.9399.2895.25
Recall10099.9599.7496.8176.6792.9593.3656.2399.3789.34
F-score10099.9399.8796.7583.6777.5191.1065.3399.3292.20

LSTM with two layerTN98.6599.8699.4191.7399.8998.6699.8699.5093.7399.89
FP1.350.140.598.270.111.340.140.506.270.11
FN8.527.3647.460.5912.218.323.5043.660.5011.20
Accuracy98.4699.7997.5898.799.6798.4799.8397.8298.9699.69
Precision65.1585.1978.3999.1693.365.3186.1982.0199.3693.53
Recall91.4892.6452.5499.4187.7991.6896.5056.3499.5088.80
F-score76.0788.7462.8799.2990.4476.2891.0566.8099.4391.10

GRU with one layerTN98.6799.8799.391.8299.8898.7899.8999.2493.2999.93
FP1.330.130.78.180.121.220.110.766.710.07
FN9.619.4246.840.6511.8912.866.7640.940.5710.66
Accuracy98.4599.7997.598.6699.6898.4799.8397.6898.8699.75
Precision65.3286.575.6799.1793.1766.2788.4375.9399.3296.11
Recall90.3990.5853.1699.3588.1187.1493.2459.0699.4389.34
F-score75.788.4762.2399.2690.5675.2990.7766.4499.3792.60

GRU with two layerTN98.6399.8599.4694.0599.998.7999.8799.3196.4599.91
FP1.370.150.545.950.11.210.130.693.550.09
FN8.083.7543.340.5611.3115.212.7834.060.4810.36
Accuracy98.4599.8297.7998.9599.798.4199.8598.0199.2399.73
Precision65.0285.5280.999.494.1665.8886.9379.4899.6494.96
Recall91.9296.2556.6699.4488.6984.7997.2265.9499.5289.64
F-score76.1190.5466.4399.4291.3474.1591.7972.0899.5892.22