Research on social media tweet classification of rainstorm disaster based on XGBoost model

Inhalt:
In the context of the rapid development of the Internet, the online social media platform represented by Weibo has become an important channel for disseminating disaster information, effective identification and utilization of tweets with realtime disaster information is of great significance for disaster early warning and emergency management. This paper takes the tweets about "720" Henan rainstorm extracted from Weibo as data set, and uses jieba word segmentation and word2vec to carry out word vector representation for all tweets, and finally uses five machine learning algorithms to classify them, the results show that the precision, recall , f1_score and AUC_score of the XGBoost model are significantly higher than those of other classification models, so the model can be used to guide the emergency management of rainstorm disaster.