A Malware Detection Scheme via Smart Memory Forensics for Windows Devices

Table 2

Comparison of state-of-the-art predictive models on optimal feature set

Classifier	Accuracy (%)	Precision (%)	Recall (%)	F1-score (%)	Loss (%)
Logistic reg.	89.4	89.8	89.4	87.1	N/A
Naïve bayes	80.4	85.4	80.4	82.2	N/A
K-near neighbor	94.8	94.7	94.8	94.5	N/A
Decision tree	87.1	86.8	87.1	86.9	N/A
Random forest	91.5	91.4	91.5	90.4	N/A
DNN	84.7	72.0	85.0	78.0	43.0
GRU	94.7	95.0	95.0	95.0	20.0
RNN	85.9	91.0	86.0	87.0	36.0
LSTM	94.5	94.0	95.0	94.0	22.0
Proposed model	98.1	98.0	98.0	98.0	10.0