Datasets:
Tasks:
Text Classification
Formats:
csv
Languages:
Portuguese
Size:
10K - 100K
Tags:
hate-speech-detection
License:
victoriadreis
commited on
Commit
•
e1df824
1
Parent(s):
c7d2106
Update README.md
Browse files
README.md
CHANGED
@@ -44,5 +44,34 @@ root.
|
|
44 |
## Voting process
|
45 |
To generate the binary matrices, we employed a straightforward voting process. Three distinct evaluations were assigned to each document. In cases where a document received two or more identical classifications, the adopted value is set to 1; otherwise, it is marked as 0.
|
46 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
## Acknowledge
|
48 |
The TuPy project is the result of the development of Felipe Oliveira's thesis and the work of several collaborators. This project is financed by the Federal University of Rio de Janeiro ([UFRJ](https://ufrj.br/)) and the Alberto Luiz Coimbra Institute for Postgraduate Studies and Research in Engineering ([COPPE](https://coppe.ufrj.br/)).
|
|
|
44 |
## Voting process
|
45 |
To generate the binary matrices, we employed a straightforward voting process. Three distinct evaluations were assigned to each document. In cases where a document received two or more identical classifications, the adopted value is set to 1; otherwise, it is marked as 0.
|
46 |
|
47 |
+
## Languages
|
48 |
+
The language used in the dataset is Brazilian Portuguese. The associated BCP-47 code is pt-BR.
|
49 |
+
|
50 |
+
# Dataset Structure
|
51 |
+
|
52 |
+
## Data Instances
|
53 |
+
|
54 |
+
A data point comprises the tweet text (a string) along with thirteen categories, each category is assigned a value of 0 when there is an absence of aggressive or hateful content and a value of 1 when such content is present. These values represent the consensus of annotators regarding the presence of aggressive, hate, ageism, aporophobia, body shame, capacitism, lgbtphobia, political, racism, religious intolerance, misogyny, xenophobia, and others.
|
55 |
+
An illustration from the multilabel ToLD-Br dataset is depicted below:
|
56 |
+
|
57 |
+
```
|
58 |
+
{'text': 'e tem pobre de direita imbecil que ainda defendia a manutenção da política de preços atrelada ao dólar link'
|
59 |
+
'aggressive': 1
|
60 |
+
'hate': 1
|
61 |
+
'ageism': 0
|
62 |
+
'aporophobia': 1
|
63 |
+
'body shame': 0
|
64 |
+
'capacitism': 0
|
65 |
+
'lgbtphobia': 0
|
66 |
+
'political': 1
|
67 |
+
'racism' : 0
|
68 |
+
'religious intolerance' : 0
|
69 |
+
'misogyny' : 0
|
70 |
+
'xenophobia' : 0
|
71 |
+
'other' : 0
|
72 |
+
}
|
73 |
+
```
|
74 |
+
## Data Fields
|
75 |
+
|
76 |
## Acknowledge
|
77 |
The TuPy project is the result of the development of Felipe Oliveira's thesis and the work of several collaborators. This project is financed by the Federal University of Rio de Janeiro ([UFRJ](https://ufrj.br/)) and the Alberto Luiz Coimbra Institute for Postgraduate Studies and Research in Engineering ([COPPE](https://coppe.ufrj.br/)).
|