Datasets:
Tasks:
Text Classification
Modalities:
Text
Sub-tasks:
hate-speech-detection
Languages:
Portuguese
Size:
1K - 10K
Tags:
instagram
DOI:
ruanchaves
commited on
Commit
·
1164bad
1
Parent(s):
dbb6556
Update README.md
Browse files
README.md
CHANGED
@@ -19,15 +19,15 @@ size_categories:
|
|
19 |
|
20 |
### Dataset Summary
|
21 |
|
22 |
-
|
23 |
|
24 |
### Supported Tasks and Leaderboards
|
25 |
|
26 |
-
|
27 |
|
28 |
### Languages
|
29 |
|
30 |
-
|
31 |
|
32 |
## Dataset Structure
|
33 |
|
@@ -37,11 +37,14 @@ size_categories:
|
|
37 |
|
38 |
### Data Fields
|
39 |
|
40 |
-
|
|
|
|
|
|
|
41 |
|
42 |
### Data Splits
|
43 |
|
44 |
-
|
45 |
|
46 |
## Dataset Creation
|
47 |
|
|
|
19 |
|
20 |
### Dataset Summary
|
21 |
|
22 |
+
HateBR is the first large-scale expert annotated corpus of Brazilian Instagram comments for hate speech and offensive language detection on the web and social media. The HateBR corpus was collected from Brazilian Instagram comments of politicians and manually annotated by specialists. It is composed of 7,000 documents annotated according to three different layers: a binary classification (offensive versus non-offensive comments), offensiveness-level (highly, moderately, and slightly offensive messages), and nine hate speech groups (xenophobia, racism, homophobia, sexism, religious intolerance, partyism, apology for the dictatorship, antisemitism, and fatphobia). Each comment was annotated by three different annotators and achieved high inter-annotator agreement. Furthermore, baseline experiments were implemented reaching 85% of F1-score outperforming the current literature models for the Portuguese language. Accordingly, we hope that the proposed expertly annotated corpus may foster research on hate speech and offensive language detection in the Natural Language Processing area.
|
23 |
|
24 |
### Supported Tasks and Leaderboards
|
25 |
|
26 |
+
Hate Speech Detection, Hate
|
27 |
|
28 |
### Languages
|
29 |
|
30 |
+
Portuguese
|
31 |
|
32 |
## Dataset Structure
|
33 |
|
|
|
37 |
|
38 |
### Data Fields
|
39 |
|
40 |
+
* instagram_comments: Instagram comments.
|
41 |
+
* offensive_language: Offensive language classification divided into offensive comments versus non-offensive comments.
|
42 |
+
* offensiveness_levels: Offensiveness-level classification divided into highly offensive, moderately offensive, and slightly offensive.
|
43 |
+
* hate_speech: Hate speech classification divided into nine different hate groups: antisemitism, apology for the dictatorship, fatphobia, homophobia, partyism, racism, religious intolerance, sexism, and xenophobia. At last, offensive & no hate speech comments also was classified.
|
44 |
|
45 |
### Data Splits
|
46 |
|
47 |
+
No standard splits have been provided by the authors.
|
48 |
|
49 |
## Dataset Creation
|
50 |
|