Twitter is one of the most common social networking platforms, and millions of tweets are generated every hour. Various non-profit organizations and relief agencies monitor Twitter data to control and help the people in emergency and need. The popularity and accessibility of smartphone to people has made this scenario possible. It allows the user to announce the emergency they are facing in real time. This paper proposes a novel framework incorporating a deep learningDeep Learning architecture transformerTransformers to predict the tweets that signify the disaster situation. The disaster dataset based on Twitter is collected, preprocessing the data. The preprocessing steps include tokenization and lemmatization, stop word removal, and word entity recognition. The individual informative tweets are extracted and enriched using semanto sim and Twitter semantic similarityTwitter Semantic Similarity after preprocessing the dataset. Disaster ontology is generated using ontocolabOntocolab, and it is incorporated with the enriched words to generate metadataMetadata for the disaster dataset.
TPredDis: Most Informative Tweet Prediction for Disasters Using Semantic Intelligence and Learning Hybridizations
Lect. Notes Electrical Eng.
International Conference on Robotics, Control, Automation and Artificial Intelligence ; 2022 November 24, 2022 - November 26, 2022
2023-11-18
10 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Online Contents | 2010
Fake Tweet Data Analysis using Machine Learning Methods
IEEE | 2021
|Etude technique: Peugeot 125 Tweet, Peugeot 125 Tweet RS
Kraftfahrwesen | 2012
|