Survey on Astroturfing Detection and Analysis from an Information Technology Perspective

Table 4

Comparison of several literature under different evaluation criterions.

Literature	Evaluation criterion
	Precision	Recall	F1	AUROC	AUPR	TPR	FPR	FNR	Accuracy	ER
Lee et al. [21]	—	—	—	—	—	0.98	0.010	—	—	—
Lee et al. [23]	—	—	0.974	—	—	—	0.008	0.248	97.35%	—
Dong et al. [24]	96.08%	94.15%	0.951	—	—	—	—	—	95.85%	—
Li et al. [25]	61.40%	62.10%	0.631	—	—	—	—	—	—	—
Liu et al. [26]	—	—	0.856	—	—	—	—	—	—	—
Aghakhani et al. [27]	—	—	—	—	—	—	—	—	89.10%	—
Xu et al. [28]	—	—	—	—	—	0.94	0.096	—	—	—
Dong et al. [29]	87.15%	83.02%	0.850	—	—	—	—	—	—	—
Fakhraei et al. [34]	>50.00%	80.00%	—	0.914 ± 0.001	0.543 ± 0.005	—	—	—	—	—
Zhang et al. [35]	—	—	0.866	—	—	—	—	—	88.15%	—
Ratkiewicz et al. [36]	—	—	—	—	—	—	—	—	96.40%	—
Liu et al. [38]	82.00%	80.00%	0.810	—	—	—	—	—	—	—
Noekhah et al. [39]	—	—	—	—	—	0.94	0.018	0.058	96.17%	—
Hu et al. [43]	86.50%	93.90%	0.901	—	—	—	—	—	—	—
Hu et al. [44]	91.30%	94.40%	0.928	—	—	—	—	—	—	—
Liu et al. [46]	79.48%	79.49%	0.793	—	—	—	—	—	78.62%	—
You et al. [47]	72.80%	73.30%	0.730	—	—	—	—	—	73.00%	—
Barushka et al. [49]	—	—	0.959	—	—	—	0.012	0.127	96.16%	—
Lee et al. [50]	—	—	0.966	—	—	—	0.036	0.174	93.26%	—
Sun et al. [51]	—	—	—	—	—	0.86	0.250	—	—	21.60%
Li et al. [53]	84.40%	95.50%	0.861	—	—	—	—	—	83.50%	—

Bold indicates the maximum value under each evaluation criterion.