基于大数据挖掘技术的地震舆情感知研究

对地震舆情信息的深入感知和有效管理,能够保障社会和谐发展。提出一个基于大数据技术和深度学习的地震舆情感知平台,基于Hadoop和MongoDB大数据技术实现对海量实时地震舆情数据的处理和存储。基于Word2vec和LSTM的融合模型能够有效实现震后网民的情感识别,为舆情预警提供支持。以台湾5.8级地震舆情数据为例,对该平台进行了验证。实验结果表明,该平台能够有效对海量地震舆情信息进行监测和分析,对地震舆情情感识别精确率达到93.76%;在准确率和收敛速度性能上均高于传统CNN神经网络,能够为后续谣言的准确识别提供有力的支持。本研究可为相关灾害类舆情研究提供一些思路。

Research on Earthquake Public Opinion Perception with Big Data Mining Technology

The in-depth perception and effective management of earthquake public opinion information can ensure the harmony in social development.In this paper,we designed a earthquake public opinion perception platform based on big data technology and deep learning.Big data technologies of Hadoop and MongoDB are adopted to process and store massive real-time seismic public opinion data.The fusion model based on Word2vec and LSTM can effectively realize netizens post-earthquake emotion recognition,and provide support for public opinion warning.This platform is validated by the public opinion data of Taiwan M 5.8 earthquake.As the results show,the platform can effectively monitor massive public opinion information of earthquake,and its precision rate of emotion recognition of earthquake public opinion reaches 93.76%.This platform is better than the traditional CNN neural network in both accuracy and convergence speed performance,which can provide strong support for subsequent rumor identification.The results in this paper could offer some ideas for related researches on disaster public opinion.