Research Article

An Optimized Hybrid Deep Learning Model to Detect COVID-19 Misleading Information

Table 9

The performance of applying ML to dataset 2.

ModelsFeature selection methodsCross-validation performanceTest performance
ACCPRERECF1ACCPRERECF1

DTUnigram96.3796.096.3896.0295.6595.1695.6595.36
Bi-gram95.8694.9895.8294.6694.9893.6294.9894.09
Tri-gram95.6594.9595.6293.995.6294.6895.6293.96
Four-gram95.594.1695.593.4495.3994.195.3993.29

KNNUnigram95.9496.0295.9494.495.4495.6595.4493.9
Bi-gram95.5595.1995.5593.5595.5895.295.5893.7
Tri-gram95.5595.1995.5593.5595.6494.9995.6493.9
Four-gram95.5395.0395.5393.595.3094.295.3093.20

LRUnigram97.2797.1297.2796.8396.6996.3896.6996.02
Bi-gram95.9195.995.9194.3795.394.2995.394.63
Tri-gram95.6695.7195.6693.895.5794.4195.5793.94
Four-gram95.6495.5595.6493.7495.094.095.093.0

RFUnigram96.9597.0196.9796.2496.2296.2896.2294.98
Bi-gram95.695.0795.5993.6695.6495.7895.6493.79
Tri-gram95.5294.5695.5193.4895.4995.5895.4993.43
Four-gram95.5194.3195.4993.4595.3590.9195.3593.07

SVMUnigram97.496.9897.096.3396.9096.8096.9096.04
Bi-gram97.096.8097.396.3496.7896.6996.7896.02
Tri-gram95.7395.6795.7393.9795.6194.8995.6193.88
Four-gram95.5795.3595.5793.5995.3994.195.3993.29

NBUnigram96.2295.6996.2295.8195.3894.5595.3894.85
Bi-gram95.3690.9395.3693.0995.3590.9195.3593.07
Tri-gram95.3690.9395.3693.0995.3590.9195.3593.07
Four-gram95.0090.0095.0093.0095.0290.0495.0193.03