Sentiment Analysis of Epidemic Comments Based on RoBERTa-WWM-RCNN
LI Hao
ZHANG Jie
Abstract：In order to help the government better understand the people's main emotional tendencies during the epidemic peri-od and make more scientific and effective epidemic prevention decisions,this paper proposes an emotional analysis model of epi-demic comment based on RoBERTa-WWM-RCNN.Firstly,the data set is preprocessed by filtering methods such as text segmenta-tion and stoplist.Secondly,RoBERTa model and whole word mask strategy are used to solve the problem of polysemy.Then,circu-lar convolution neural network model is used to obtain more comprehensive contextual semantic feature information.Finally,the emotional polarity of comment data is obtained by Softmax classifier,including positive,neutral and negative.The experimental re-sults show that the model has the best effect on the data set of Weibo epidemic review,with the accuracy rate of 75.62%and the Macro_F1 value of 72.05%.
Keywords：deep learningepidemic sentiment analysisRoBERTa-WWM modelRCNN model
Publication Date：2025-10-20
Online Publishing Date：2026-01-16(First online date of this platform, not the publication date of the document)
Pages：6( 2708-2712,2727 )