基于BERT和LSTM的地震舆情情感倾向分析  

Earthquake public opinion sentiment tendency analysis based on BERT and LSTM

作　　者：吴月波 刘克辉 石晓辉 WU Yue-bo;LIU Ke-hui;SHI Xiao-hui

机构地区：[1]山东省地震局烟台地震监测中心站,山东烟台264001

出　　处：《智能城市》2024年第10期27-29,共3页Intelligent City

基　　金：中国地震局地震应急青年重点任务(CEA_EDEM-20240208)。

摘　　要：通过Scrapy框架获取近三年(2021—2023年)国内5.0级以上地震震后48 h的相关微博博文和评论,对爬取的文本数据进行预处理并组成数据集,设计基于BERT和长短时记忆(LSTM)网络的深度学习地震舆情情感倾向模型。结果显示,该模型在地震舆情文本情感分析的准确率达到97.8%,具有高效的特征提取能力,能够为地震网络舆情监测提供参考。Through the Scrapy framework,the relevant Weibo blog posts and comments related to earthquake of magnitude 5.0 or above in China in the past three years(2021-2023)were obtained,and the crawled text data was preprocessed and compiled into a dataset,and a deep learning earthquake public opinion sentiment tendency model based on BERT and Long Short-Term Memory(LSTM)was designed.The results show that the accuracy of the model in the text sentiment analysis of earthquake public opinion reaches 97.8%,and it has efficient feature extraction ability,which can provide a reference for the monitoring of earthquake network public opinion.

关 键 词：Scrapy框架 地震舆情 BERT LSTM 情感分析 

分 类 号：P315[天文地球—地震学]