Table 2.
Dataset | Bot predictions | Hate scores | Offensive scores | |||
---|---|---|---|---|---|---|
Tweets | Users | Tweets | Users | Tweets | Users | |
US |
3.026 M (26.31%) |
237 K (14.91%) | 0.1023 (0.0769) | 0.1073 (0.0656) |
0.2773 (0.1519) |
0.2914 (0.1418) |
PH | 3.436 M (21.73%) | 150 K (15.70%) | 0.0896 (0.0717) | 0.0924 (0.0600) |
0.2672 (0.1375) |
0.2836 (0.1180) |
For this table, an 80% probability threshold was used to classify a user as a bot. Bot tweets refer to tweets produced by bots predicted by the same threshold. Parentheses provide dataset proportions for BotHunter predictions and standard deviations for hate speech and offensive speech scores