metadata

license: cc-by-nc-sa-4.0
task_categories:
  - text-classification
language:
  - tr
tags:
  - text classification
  - event detection
  - topic detection
  - tweets
  - social media
  - topic classification
size_categories:
  - n<1K

Turkish Event Detection Tweet Dataset: BilTweetNews

The dataset contains tweets related to six major events from Turkish news sources between May 4, 2015 and Jan 8, 2017.

There are 7 event classes:

E1: May 25, 2015 One of the popular football clubs in Turkey, Galatasaray, wins the 2015 Turkish Super League.
E2: Sep 6, 2015 A terrorist group, called PKK, attacked to soldiers in Dağlıca, a village in southeastern Turkey.
E3: Oct 7, 2015 A Turkish scientist, Aziz Sancar, won the 2015 Nobel Chemistry prize with his studies on DNA repair.
E4: May 27, 2016 A local football club of Alanya promoted to the Turkish Super League for the first time in their history.
E5: Jun 17, 2016 A traditional anthem that is mostly played by secularists in Turkey, called the 10th Year Anthem, was forbidden in schools by the director of national education in the Black Sea province of Bolu.
E6: Oct 17, 2016 A magazine programmer confused that Madonna in a Fur Coat, a book written in 1943 by a Turkish celebrated writer, Sabahattin Ali, was about popstar Madonna’s life. The book tells a story between a Turkish student and German singer after the World War I.
Other: Not related to any news topic

For each event, 100 related-candidate and 60 unrelated-candidate tweets are selected. Lastly, we randomly select 40 tweets that are potentially not related at all, 5 of them are removed due to detecting near-duplicates later. The dataset has 995 tweets in total.

The task of this dataset is event detection. The sentiment analysis labels can be found at https://huggingface.co/datasets/ctoraman/BilTweetNews-Sentiment

All tweets are labeled by 17 annotators. We provide the normalized distribution of annotations across 7 event classes. We also provide the majority class at the last column. There are no cases where multiple classes have the highest score.

Github Repo: https://github.com/BilkentInformationRetrievalGroup/BilTweetNews2017

If you would like to use any material in this repository, please cite the following papers:

Toraman, C. Early Prediction of Public Reactions to News Events Using Microblogs. Seventh BCS-IRSG Symposium on Future Directions in Information Access (FDIA 2017), Barcelona, Spain, 5 September 2017.
Toraman, C. Event-related microblog retrieval in Turkish. Turkish Journal of Electrical Engineering and Computer Sciences. 2021. DOI: 10.3906/elk-2108-167