Classification of Events Tweets Using Machine Learning
Tarun Kumar Jain, Dinesh Gopalani, Yogesh Kumar Meena · 2023
Microblogging platforms like Twitter have turned into a significant correspondence divert in the midst of crisis. The omnipresence of cell phones empowers individuals to declare a crisis they’re seeing continuously. Along these lines, more offices are keen on programmatically observing Twitter (for example calamity help associations and news organizations). Be that as it may, it’s not commonly apparent whether a singular’s words are truly pronouncing a catastrophe. Because of this many reserachers are working to use this fact of information and utilize the same. So here in this research study we have prepared and tried 9 classifiers with 3 different feature extraction technique and investigated their exhibition. That’s what the proposed exploratory outcomes showed, Transformer based architecture like BERT yield best execution with the most noteworthy exactness of 83.92%, trailed by the Ensemble voting classifier with a F1-Score of 77.90% and SVM classifier with TF-IDF at 79.41%.