ClickbaitTR: Dataset for clickbait detection from Turkish news sites and social media with a comparative analysis via machine learning algorithms

Genc S., SÜRER E.

JOURNAL OF INFORMATION SCIENCE, cilt.49, sa.2, ss.480-499, 2023 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 49 Sayı: 2
  • Basım Tarihi: 2023
  • Doi Numarası: 10.1177/01655515211007746
  • Dergi Adı: JOURNAL OF INFORMATION SCIENCE
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Scopus, Academic Search Premier, FRANCIS, IBZ Online, Periodicals Index Online, ABI/INFORM, Aerospace Database, Analytical Abstracts, Applied Science & Technology Source, Business Source Elite, Business Source Premier, Communication Abstracts, Compendex, Computer & Applied Sciences, EBSCO Education Source, Education Abstracts, Index Islamicus, Information Science and Technology Abstracts, INSPEC, Library and Information Science Abstracts, Library Literature and Information Science, Library, Information Science & Technology Abstracts (LISTA), Metadex, Civil Engineering Abstracts
  • Sayfa Sayıları: ss.480-499
  • Anahtar Kelimeler: clickbait detection, Data analysis, dataset formation, neural networks
  • Orta Doğu Teknik Üniversitesi Adresli: Evet

Özet

Clickbait is a strategy that aims to attract people's attention and direct them to specific content. Clickbait titles, created by the information that is not included in the main content or using intriguing expressions with various text-related features, have become very popular, especially in social media. This study expands the Turkish clickbait dataset that we had constructed for clickbait detection in our proof-of-concept study, written in Turkish. We achieve a 48,060 sample size by adding 8859 tweets and release a publicly available dataset - ClickbaitTR - with its open-source data analysis library. We apply machine learning algorithms such as Artificial Neural Network (ANN), Logistic Regression, Random Forest, Long Short-Term Memory Network (LSTM), Bidirectional Long Short-Term Memory (BiLSTM) and Ensemble Classifier on 48,060 news headlines extracted from Twitter. The results show that the Logistic Regression algorithm has 85% accuracy; the Random Forest algorithm has a performance of 86% accuracy; the LSTM has 93% accuracy; the ANN has 93% accuracy; the Ensemble Classifier has 93% accuracy; and finally, the BiLSTM has 97% accuracy. A thorough discussion is provided for the psychological aspects of clickbait strategy focusing on curiosity and interest arousal. In addition to a successful clickbait detection performance and the detailed analysis of clickbait sentences in terms of language and psychological aspects, this study also contributes to clickbait detection studies with the largest clickbait dataset in Turkish.