Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -5,5 +5,34 @@ task_categories:
|
|
5 |
language:
|
6 |
- en
|
7 |
size_categories:
|
8 |
-
-
|
9 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
language:
|
6 |
- en
|
7 |
size_categories:
|
8 |
+
- 1K<n<200K
|
9 |
+
---
|
10 |
+
|
11 |
+
### Dataset info
|
12 |
+
#### Training Dataset:
|
13 |
+
You are provided with a large number of Wikipedia comments which have been labeled by human raters for toxic behavior. The types of toxicity are:
|
14 |
+
|
15 |
+
- toxic
|
16 |
+
- severe_toxic
|
17 |
+
- obscene
|
18 |
+
- threat
|
19 |
+
- insult
|
20 |
+
- identity_hate
|
21 |
+
|
22 |
+
The original dataset can be found here: [jigsaw_toxic_classification](https://www.kaggle.com/competitions/jigsaw-toxic-comment-classification-challenge/data)
|
23 |
+
|
24 |
+
Our training dataset is a version from the original dataset, <b>containing equal number of samples from each classes. </b><br>
|
25 |
+
Number of samples in train_dataset: 25960<br>
|
26 |
+
Number of samples in val_dataset: 6490<br>
|
27 |
+
Number of samples in test_dataset: 153164<br>
|
28 |
+
|
29 |
+
|
30 |
+
### Citation
|
31 |
+
<pre>
|
32 |
+
@misc{jigsaw-toxic-comment-classification-challenge,
|
33 |
+
author = {cjadams, Jeffrey Sorensen, Julia Elliott, Lucas Dixon, Mark McDonald, nithum, Will Cukierski},
|
34 |
+
title = {Toxic Comment Classification Challenge},
|
35 |
+
publisher = {Kaggle},
|
36 |
+
year = {2017},
|
37 |
+
url = {https://kaggle.com/competitions/jigsaw-toxic-comment-classification-challenge}
|
38 |
+
}</pre>
|