Overview of text classification for emergency related information on social platforms
Jiang Yuqi, Qiang Zishan, Bu Fanliang
College of Information Network Security, People′s Public Security University of China
Abstract： In the early stages of an emergency event, timely extraction of valuable information from massive social media data holds great significance in providing decision-making references for emergency response. With the rapid development of natural language processing, text classification has gradually been applied in this field, mainly divided into traditional machine learning based methods such as K-Nearest Neighbor, Naive Bayes, Decision Tree, Support Vector Machines, and deep learning based methods such as CNN, RNN, GCN and Transformer. This paper analyzes the current mainstream text classification methods from aspects including algorithm principles, development history, applicable fields, advantages and disadvantages. It investigates the research status and hotspots of text classification for emergency-related information on social media platforms, summarizes the problems and challenges faced by existing methods, and presents future research directions, providing references and inspiration for subsequent scientific research work.
Key words : text classification; machine learning; deep learning; emergency related information on social platforms
