Sefara, Tshephisho JRangata, Mapitsi R2023-10-132023-10-132023-10Sefara, T.J. & Rangata, M.R. 2023. A natural language processing technique to identify exaggerated news titles. http://hdl.handle.net/10204/13147 .2367-3370https://doi.org/10.1007/978-981-99-5166-6_64http://hdl.handle.net/10204/13147Exaggerated news titles are used to deceive news readers and spread misleading information. This paper presents a new natural language processing (NLP) technique that identifies exaggerated news titles. The technique uses Jaccard similarity as a pre-processing step to filter out unrelated articles. The technique then applies text summarisation on the content of the news article to create a new title. Lastly, the technique applies cosine similarity to compare similar articles between the article title and the newly generated titles. The output is the classification of the news articles using the output of cosine similarity. This technique performed well in major South African news articles.FulltextenText similarityText summarisationNatural Language ProcessingCosine similarityJaccard similarityConference PresentationSefara, T. J., & Rangata, M. R. (2023). A natural language processing technique to identify exaggerated news titles. http://hdl.handle.net/10204/13147Sefara, Tshephisho J, and Mapitsi R Rangata. "A natural language processing technique to identify exaggerated news titles." <i>Lecture Notes in Networks and Systems, 757 (paper presented at ICICCT 2023: Inventive Communication and Computational Technologies)</i> (2023): http://hdl.handle.net/10204/13147Sefara TJ, Rangata MR, A natural language processing technique to identify exaggerated news titles; 2023. http://hdl.handle.net/10204/13147 .TY - Conference Presentation AU - Sefara, Tshephisho J AU - Rangata, Mapitsi R AB - Exaggerated news titles are used to deceive news readers and spread misleading information. This paper presents a new natural language processing (NLP) technique that identifies exaggerated news titles. The technique uses Jaccard similarity as a pre-processing step to filter out unrelated articles. The technique then applies text summarisation on the content of the news article to create a new title. Lastly, the technique applies cosine similarity to compare similar articles between the article title and the newly generated titles. The output is the classification of the news articles using the output of cosine similarity. This technique performed well in major South African news articles. DA - 2023-10 DB - ResearchSpace DP - CSIR J1 - Lecture Notes in Networks and Systems, 757 (paper presented at ICICCT 2023: Inventive Communication and Computational Technologies) KW - Text similarity KW - Text summarisation KW - Natural Language Processing KW - Cosine similarity KW - Jaccard similarity LK - https://researchspace.csir.co.za PY - 2023 SM - 2367-3370 T1 - A natural language processing technique to identify exaggerated news titles TI - A natural language processing technique to identify exaggerated news titles UR - http://hdl.handle.net/10204/13147 ER -27163