Skip to content

Public Behavior Analysis under the COVID-19 Emergency——Based on Weibo Mining

Notifications You must be signed in to change notification settings

hyliush/COVID-19-Public-behavior-sentiment-and-attention

Repository files navigation

Public behavior sentiment and attention Public Behavior Analysis under the COVID-19 Emergency——Based on Weibo Mining

疫情期间完成的一些工作,时间紧张,自己也没这方面的研究经验,没有投出去。大家感兴趣的话,可以参考一下。 NLP部分的模型不是特别复杂,都是一些传统常用的模型 数据见https://github.com/nghuyong/weibo-public-opinion-datasets

【方法/过程】

本文利用机器学习有监督地识别得到2019年12月31日至2020年2月23日期间来自345个城市244,880条关于新冠肺炎的微博数据,从中提取公众情感指数和关注指数,讨论了疫情发展四个阶段全国范围内公众的情感和关注程度变化,进一步对比分析不同阶段不同地区的差异,最后对公众行为的阶段差异性,以及和经济发展水平、相距重灾区距离的相关性进行了检验。

【Method/process】

This paper used machine learning to identify 244,880 microblog posts from 345 cities from December 31, 2019 to February 23, 2020, extracted public sentiment and focus index and discussed the nation's overall emotional trend and the degree of attention to this event in the four stages of the development of the epidemic. It further analyzed the differences in public behaviors in different regions at different stages, and finally the correlation between the public behavior, the level of economic development and the distance from the severely affected area was tested.

【结果/结论】

公众情绪经历了持续波动、上升,稳定正面三种状态;公众对新冠肺炎疫情的关注趋势表现为低关注、关注波动、逐渐上升和逐渐下降四个阶段;公众情绪在不同阶段均存在显著差异,对疫情的关注度在部分阶段存在差异;公众行为和区域位置相关,表现为公众所在地与重灾区接近程度和情感指数呈显著负相关,所在地经济发展水平和关注指数呈显著正相关。

【Result/conclusion】

The study found that public sentiment experienced three states: continuous fluctuation, rising, and stable positive; the public’s attention to the trend of the Novel coronavirus pneumonia showed four stages: low attention, fluctuation of concern, gradual rise, and gradual decline; there are significant differences in public sentiment in different stages, and the degree of attention to the epidemic differs in some stages; public behavior is related to regional location, which was shown by the close negative correlation between distance from the severely affected area and the sentiment index, and the level of economic development of the location and the focus index showed a significant positive correlation.

走势图

词频统计

区域差异

Citation

If you use this work in a scientific publication, I would appreciate that you can also cite the following BibTex entry:

@misc{hyliu2020@COVID-19-Public-behavior-sentiment-and-attention, title=Public-behavior-sentiment-and-attention-Public-Behavior-Analysis-under-the-COVID-19-Emergency——Based on Weibo Mining}, author={Hengyuan Liu}, howpublished={\url{https://github.com/hyliush/COVID-19-Public-behavior-sentiment-and-attention}}, year={2020} }

For any support, queries, contact hyliu_sh@outlook.com

About

Public Behavior Analysis under the COVID-19 Emergency——Based on Weibo Mining

Resources

Stars

Watchers

Forks

Packages

No packages published