Table 2 Detection results for cluster feature combinations
From: An unsupervised method for social network spammer detection based on user information interests
Dataset | Features | Accuracy | Precision | Recall |
|---|---|---|---|---|
Social Honeypot | GOSS + LOSS | 0.934 | 0.92 | 0.91 |
GOSS + LOSS + Entropy | 0.967 | 0.96 | 0.95 | |
GOSS + Entropy | 0.927 | 0.91 | 0.90 | |
LOSS + Entropy | 0.950 | 0.94 | 0.93 | |
The Fake Project | GOSS + LOSS | 0.954 | 0.94 | 0.93 |
GOSS + LOSS + Entropy | 0.969 | 0.95 | 0.96 | |
GOSS + Entropy | 0.947 | 0.94 | 0.91 | |
LOSS + Entropy | 0.955 | 0.95 | 0.93 | |
HSpam14 | GOSS + LOSS | 0.927 | 0.91 | 0.88 |
GOSS + LOSS + Entropy | 0.949 | 0.92 | 0.93 | |
GOSS + Entropy | 0.923 | 0.91 | 0.86 | |
LOSS + Entropy | 0.937 | 0.92 | 0.89 |