Skip to main content

Table 3 Offensive language detection—OLD data statistics [44]

From: Social media text analytics of Malayalam–English code-mixed using deep learning

Class

Train

Valid

Test

Not offensive

14,153 (88.4%)

1779 (88.99%)

1770 (88.5%)

Not-Malayalam

1287 (8.03%)

163 (8.15%)

161 (8.04%)

Offensive Targeted Insult Individual

239 (1.49%)

24 (1.20%)

29 (1.44%)

Offensive Untargeted

191 (1.19%)

20 (1.00%)

24 (1.19%)

Offensive Targeted Insult Group

140 (0.87%)

13 (0.65%)

17 (0.84%)

Total

16 010

1999

2001