Alienmaster commited on
Commit
40661d1
1 Parent(s): 9d9faf1

Detail fixed

Browse files

Changed dataformat and added details.

README.md CHANGED
@@ -4,19 +4,45 @@ language:
4
  pretty_name: "Amharic Hate Speech Dataset"
5
  tags:
6
  - am
 
 
7
  task_categories:
8
  - text-classification
9
  ---
10
 
11
- [GitHub](https://github.com/uhh-lt/AmharicHateSpeech)
12
-
13
  # Introduction
14
- The Amharic Hate Speech data is collected using the Twitter API spanning from October 1, 2020 - November 30, 2022, considering the socio-political dynamics of Ethiopia in Twitter space. We used [WEbAnno](http://ltdemos.informatik.uni-hamburg.de/codebookanno-cba/) tool for data annotation; each tweet is annotated by two native speakers and curated by one more experienced adjudicator to determine the gold labels. A total of 15.1k tweets consisting of three class labels namely: Hate, Offensive and Normal are presented. Read our papers for more details about the dataset (see below).
 
15
 
16
  # Amharic Hate Speech Data Annotation: Lab-Controlled Annotation
17
- The dataset is annotated by two annotators and a curator to determine the gold labels.
 
 
 
18
 
 
19
  For more details, You can read our paper entitled:
20
- 1. [Exploring Amharic Hate Speech data Collection and Classification Approaches](https://www.inf.uni-hamburg.de/en/inst/ab/lt/publications/2023-ayele-et-al-hate-ranlp.pdf)
21
 
 
22
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  pretty_name: "Amharic Hate Speech Dataset"
5
  tags:
6
  - am
7
+ size_categories:
8
+ - 10K<n<100K
9
  task_categories:
10
  - text-classification
11
  ---
12
 
 
 
13
  # Introduction
14
+ The Amharic Hate Speech data is collected using the Twitter API spanning from October 1, 2020 - November 30, 2022, considering the socio-political dynamics of Ethiopia in Twitter space. We used [WebAnno](http://ltdemos.informatik.uni-hamburg.de/codebookanno-cba/) tool for data annotation; each tweet is annotated by two native speakers and curated by one more experienced adjudicator to determine the gold labels. A total of 15.1k tweets consisting of three class labels namely: Hate, Offensive and Normal are presented. Read our papers for more details about the dataset (see below).
15
+
16
 
17
  # Amharic Hate Speech Data Annotation: Lab-Controlled Annotation
18
+ The dataset is annotated by two annotators and a curator to determine the gold labels. The annotation guideline can be found [here](https://github.com/uhh-lt/AmharicHateSpeech/blob/main/Data/RANLP2023/Annotation%20Guideline.pdf)
19
+
20
+ # Dataset Details
21
+ The prefix `__label__` has been removed from the labels in the uploaded version. When using the training script linked below the prefix must be added manually.
22
 
23
+ # Citation Information and Links
24
  For more details, You can read our paper entitled:
25
+ 1. [Exploring Amharic Hate Speech data Collection and Classification Approaches](https://aclanthology.org/2023.ranlp-1.6/)
26
 
27
+ Or visit our [GitHub Repository](https://github.com/uhh-lt/AmharicHateSpeech) for the papers, models and training code.
28
 
29
+ ```
30
+ @inproceedings{ayele-etal-2023-exploring,
31
+ title = "Exploring {A}mharic Hate Speech Data Collection and Classification Approaches",
32
+ author = "Ayele, Abinew Ali and
33
+ Yimam, Seid Muhie and
34
+ Belay, Tadesse Destaw and
35
+ Asfaw, Tesfa and
36
+ Biemann, Chris",
37
+ editor = "Mitkov, Ruslan and
38
+ Angelova, Galia",
39
+ booktitle = "Proceedings of the 14th International Conference on Recent Advances in Natural Language Processing",
40
+ month = sep,
41
+ year = "2023",
42
+ address = "Varna, Bulgaria",
43
+ publisher = "INCOMA Ltd., Shoumen, Bulgaria",
44
+ url = "https://aclanthology.org/2023.ranlp-1.6",
45
+ pages = "49--59",
46
+ abstract = "In this paper, we present a study of efficient data selection and annotation strategies for Amharic hate speech. We also build various classification models and investigate the challenges of hate speech data selection, annotation, and classification for the Amharic language. From a total of over 18 million tweets in our Twitter corpus, 15.1k tweets are annotated by two independent native speakers, and a Cohen{'}s kappa score of 0.48 is achieved. A third annotator, a curator, is also employed to decide on the final gold labels. We employ both classical machine learning and deep learning approaches, which include fine-tuning AmFLAIR and AmRoBERTa contextual embedding models. Among all the models, AmFLAIR achieves the best performance with an F1-score of 72{\%}. We publicly release the annotation guidelines, keywords/lexicon entries, datasets, models, and associated scripts with a permissive license.",
47
+ }
48
+ ```
dev.tsv ADDED
The diff for this file is too large to render. See raw diff
 
test.tsv ADDED
The diff for this file is too large to render. See raw diff
 
train.tsv ADDED
The diff for this file is too large to render. See raw diff
 
with_annotators_dev.tsv ADDED
The diff for this file is too large to render. See raw diff
 
with_annotators_test.tsv ADDED
The diff for this file is too large to render. See raw diff
 
with_annotators_train.tsv ADDED
The diff for this file is too large to render. See raw diff