[1]徐晓桐,陈吉锋,李东平,等.基于多头自注意力机制的CNN地震舆情分析模型研究[J].华南地震,2024,(04):21-32.[doi:10.13512/j.hndz.2024.04.03]
 XU Xiaotong,CHEN Jifeng,LI Dongping,et al.CNN Earthquake Public Opioion Analysis Model Based on Multi-Head Self-Attention Mechanism[J].,2024,(04):21-32.[doi:10.13512/j.hndz.2024.04.03]
点击复制

基于多头自注意力机制的CNN地震舆情分析模型研究()
分享到:

华南地震[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2024年04期
页码:
21-32
栏目:
地震科学研究
出版日期:
2024-12-30

文章信息/Info

Title:
CNN Earthquake Public Opioion Analysis Model Based on Multi-Head Self-Attention Mechanism
文章编号:
1001-8662(2024)04-0021-12
作者:
徐晓桐陈吉锋李东平吴凌杰李环宇姚 迪
浙江省地震局,杭州 310063
Author(s):
XU XiaotongCHEN JifengLI DongpingWU LingJieLI HuanyuYAO Di
Zhejiang Earthquake Agency , Hangzhou 310063, China
关键词:
深度学习地震舆情多头自注意力机制情感分类
Keywords:
Deep learningEarthquake public opinionMulti-head self-attention mechanismSentiment classifi?cation
分类号:
TP391
DOI:
10.13512/j.hndz.2024.04.03
文献标志码:
A
摘要:
随着互联网的发展进步,新媒体平台逐步成为普通公众发布和获取地震灾情信息的首选途径,成为地震相关部门迅速了解当前灾情和公众舆论环境的有效渠道之一。运用网络爬虫技术,搜集震后微博用户公开发表的博文与评论构建数据集并进行预处理,为后续分析和建模奠定基础。通过引入多头自注意力机制优化传统CNN模型,构建基于多头自注意力机制的CNN地震舆情分析模型,丰富特征子空间的多样性,并行处理以及捕捉不同级别的特征和信息,增强模型对地震舆情的理解能力。利用模型对2024年3月7日青海玉树州杂多县5.5级地震进行实例应用,对灾后舆情做了可视化展示。通过实验对比,构建的模型加权平均F1达到92.9%、宏平均F1达到92.1%,能够为地震相关部门在震后快速了解灾情情况和公众舆论环境提供辅助支撑。
Abstract:
With the advancement of the internet, new media platforms are progressively becoming the preferred channels for the general public to release and access earthquake disaster information,as well as one of the effective avenues for earthquake-related departments to promptly grasp the current disaster situation and public opinion. This paper employed web crawling technology to collect post-earthquake Weibo posts and comments from users and con?structed a dataset that was then subjected to preprocessing,thus laying the foundation for subsequent analysis and modeling. The paper introduced a multi-head self-attention mechanism to optimize the conventional CNN model, thereby developing a CNN earthquake public opinion analysis model based on the multi-head self-attention mecha?nism. The paper enriched the diversity of feature subspace,ensured parallel processing,captured different levels of features and information,and enhanced the model’s ability to understand earthquake public opinion. The model was put into practical application and visualization by analyzing the public opinion following the Zadoi M5.5 earth?quake, in Yushu Tibetan Autonomous Prefecture, Qinghai Province on March 7th, 2024. Through experimental comparisons,the constructed model achieved a weighted average F1 score of 92.9% and a macro-average F1 score of 92.1%. These results demonstrate that the model can effectively provide auxiliary support for earthquake-related departments to quickly understand the disaster situation and public opinion environment after the earthquake..

参考文献/References:

[1]周义棋,田向亮,钟茂华.基于微博数据的自然灾害应急救助需求评估[J].清华大学学报:自然科学版,2022,62 (10):10.
[2] Seki Y. Use of twitter for analysis of public sentiment for im?provement of local government service[C]//2016 IEEE Inter?national Conference on Smart Computing (SMARTCOMP). St. Louis,MO,USA:IEEE ,2016:1-3.
[3]金城,吴文渊,陈柏儒,等.面向不同用户群体的社交媒体台风舆情演化分析及对比研究[J].地球信息科学学报, 2021,23(12):2174-2186.
[4] Biswas R,Alam T,Househ M,et al. Public sentiment towards vaccination after COVID-19 outbreak in the arab world[J]. Stud Health Technol Inform,2022(289):57-60.
[5]简益波,徐磊.浙江省的涉震舆情分析与应对[J].农业灾害研究,2020,10(7):138-139+150.
[6]刘耀辉,刘婉婷,张文焯,等.基于新浪微博数据的云南漾濞6.4级地震舆情时空特征及情感分析[J].自然灾害学报, 2022,31(1):168-178.
[7]李亚芳,王新刚,梁庆云.基于新浪微博大数据的新疆伽师6.4级地震舆情分析及可视化研究[J].内陆地震,2020, 34(1):103-110.
[8] Kumar L K,Thatha V N,Udayaraju P,et al. Analyzing public sentiment on the Amazon website:A GSK-based double path transformer network approach for sentiment analysis[J]. IEEE Access,2024(12):28972-28987.
[9] Kumar V V,Sahoo A,Balasubramanian S K,et al. Mitigating healthcare supply chain challenges under disaster conditions:a holistic AI-based analysis of social media data[J/OL]. Inter?national Journal of Production Research,2024:1-19[2024-3-15].https://doi.org/10.1080/00207543.2024.2316884
[10] Rout L,Acharya M K,Acharya S. Content analysis of You?Tube videos regarding natural disasters in India and analy?sis of users sentiment through viewer comments[J]. Natural Hazards,2024,120(1):219-234.
[11]唐焕玲,窦全胜,于立萍,等.有监督主题模型的SLDA-TC文本分类新方法[J]. 电子学报,2019,47(6):1300-1308.
[12] Yang C H T,Hsu Y W,Chun C L. An information hiding method based on grouping color palette indices by K-means clustering and similar pairs[C]//2012 Sixth International Conference on Genetic and Evolutionary Computing. Kitaky?ushu,Japan:IEEE ,2012:47-50.
[13] Huang X,Jin H,Zhang Y. Risk assessment of earthquake network public opinion based on global search BP neural network[J]. PloS one,2019,14(3):1-14.
[14] Taufek T E,Nor N F M,Jaludin A,et al. Public perceptions on climate change:a sentiment analysis approach[J]. GEMA Online Journal of Language Studies,2021,21(4):209-233.
[15]袁健,董光文.多维特征融合的混合神经网络文本情感分析模型[J].小型微型计算机系统,2023,44(10):2137-2143.
[16] Xu C,Gao Y. Multi-modal transformer with multi-head at?tention for emotion recognition[C]//2023IEEE International Conference on Sensors,Electronics and Computer Engineer?ing(ICSECE). Jinzhou,China:IEEE,2023:826-831.
[17]李佳静,李盛,戴媛媛,等.融合位置信息和交互注意力的方面级情感分析[J/OL].计算机工程与应用,2024:1-10[2024-3-15]. https ://link. cnki. net/urlid/11. 2127. TP. 20240122.1411.006
[18] Vosta S,Yow K C. KianNet:A violence detection model us?ing an attention-based CNN-LSTM structure[J]. IEEE Ac?cess,2023(12):2198-2209
[19] Jung K,Kim M,Chung W. Epoch-level and sequence-level multi-head self-attention-based sleep stage classification[C]//2023 11th International Winter Conference on Brain-Com?puter Interface(BCI). Gangwon,Korea:IEEE ,2023:1-6.
[20] Yue W,Xu Y,Xiang L,et al. Prediction of 3-D ocean tem?perature based on self-attention and predictive RNN[J]. IEEE Geoscience and Remote Sensing Letters,2024(21):1-5.

相似文献/References:

[1]张燕明,张红才,陈惠芳,等.实时智能地震处理系统在2013年福建仙游ML5.0地震序列中的应用研究[J].华南地震,2023,(04):64.[doi:10.13512/j.hndz.2023.04.09]
 ZHANG Yanming,ZHANG Hongcai,CHEN Huifang,et al.Application of Real-time Intelligent Seismic Processor in Xianyou ML5.0 Earthquake Sequence in Fujian Province in 2013[J].,2023,(04):64.[doi:10.13512/j.hndz.2023.04.09]
[2]梁 明,刘 军,洪玉清,等.地震自动编目处理系统在2023年广东河源M4.3地震序列中的应用[J].华南地震,2023,(04):119.[doi:10.13512/j.hndz.2023.04.15]
 LIANG Ming,LIU Jun,Hong Yuqing,et al.Application of Seismic Automatic Cataloguing Processing System in Heyuan M4.3 Earthquake Sequence in Guangdong in 2023[J].,2023,(04):119.[doi:10.13512/j.hndz.2023.04.15]

备注/Memo

备注/Memo:
收稿日期:2024-00-00
基金项目:中国地震局重大政策理论与实践问题研究课题(CEAZY2024JZ01);中国地震局地震应急与信息青年重点任务(CEAEDEM202310)联合资助。
作者简介:徐晓桐(1995-),女,工程师,研究方向为深度学习情感分类、地震舆情分析。
通信作者:陈吉锋(1980-),男,高级工程师,研究方向为自然语言处理、地震信息化、地震舆情分析。E-mail:xuxiaotong@zjdz.gov.cn
更新日期/Last Update: 2024-12-30