摘要
[目的/意义]分析突发事件的微博情感演变态势,可以发现舆情演变规律和潜在风险,为舆情引导提供决策支持。[方法/过程]提出基于情感分析的突发事件微博舆情演变分析模型。用Python编写爬虫获取微博评论,经过数据预处理和分词,采用朴素贝叶斯分类器进行情感分析。定义情感热度,并据此将舆情演变过程划分为高热期,持续期、反复期、消亡期。运用统计和数据可视化方法研究各阶段评论词云、情感倾向演化、地域热度和各年龄段演化特点,分析舆情时空规律。[结果/结论]以天津8·12危化品爆炸事故为例,建立微博舆情演化分析模型分析舆情演变。实证分析表明,微博舆情演化分析模型可以合理划分演变阶段,发现各阶段演化规律和热点内容。
[Purpose/Significance]Analyzing the emotion evolution of the emergency is helpful for discovering the rules of evolution of public opinion and mastering the characteristics and potential risks,thus providing theoretical support for the decision-making process.[Method/Process]In this paper,a model of public opinion evolution on the microblogs about emergency events based on emotion analysis is proposed.Firstly,the crawler was written in Python language to obtain the comment data of microblogs.On the basis of data preprocessing and text segmentation,the naive Bayes classifier was adopted to conduct the emotion analysis.The definition of emotion heat indicator was put forward.The development and evolution process of public opinion of emergency events were divided into four stages:the hyperpyrexia period,the continuous focus period,the repeated public attention period and the phasing out period.Then,the statistical method and data visualization were used to study the features of comment,emotion evolution tendency,regional characteristics and age group differences in different stages.[Result/Conclusion]Taking the"8·12 Explosion Accident in Tianjin"as an example,we learned about the laws of evolution trend of public opinion.Verification shows that the analysis model of Weibo public opinion evolution can reasonably divide the evolution stages of public opinion and help grasp the evolution characteristics and laws of each stage and identify the hot issues of each stage.
作者
任中杰
张鹏
李思成
兰月新
夏一雪
崔彦琛
Ren Zhongjie;Zhang Peng;Li Sicheng;Lan Yuexin;Xia Yixue;Cui Yanchen(China People’s Police University, Langfang 065000)
出处
《情报杂志》
CSSCI
北大核心
2019年第2期140-148,共9页
Journal of Intelligence
基金
教育部人文社会科学基金"面向突发事件的网络流言风险预警及对策研究"(编号:17YJC630214)
公安部理论与软科学项目"基于网络舆情综合研判的群体性事件预警研究"(编号:2015LLYJWJXY034)
全国统计科学研究重点项目"舆情大数据环境下突发事件民意监测与评估研究"(编号:2017LZ37)
河北省科技计划项目"新媒体环境下突发事件危机信息挖掘与决策关键技术研究"(编号:17455610)
武警学院国家社科基金培育项目"大数据环境下公共危机事件情报挖掘与决策预警研究"(编号:SKJJPY201721)
关键词
微博
舆情
数据挖掘
情感
舆情演化
Weibo
public opinion
data mining
emotion
public opinion evolution
