Scholarlink

<p>In the context of digital transformation in public discourse, computational analysis of evolving hot topics in online public opinion regarding energy corporations provides critical insights for market intelligence and public opinion guidance. This study proposes an innovative framework that integrates topic modeling and multidimensional analytics to decode public perception patterns of energy corporations. Original text data related to eight renowned energy corporations were collected on the Sina Weibo platform from October 1, 2021, to September 30, 2023. By employing the Latent Dirichlet Allocation (LDA) probabilistic topic model, topic clustering was conducted, followed by a secondary dimensionality reduction of clustering results through the integration of inter-topic distances and text analysis. Subsequently, a comprehensive analysis of hot topics (themes) concerning energy corporations was conducted using co-occurrence vocabulary, a heat index, and sentiment analysis with a visual presentation of evolutionary trends. Online public opinion on energy corporations can be categorized into seven hot topics, which can be further summarized into three major themes. The content evolution of hot topics remains relatively stable, with topic generation being driven by leading corporations in new energy businesses. The popularity evolution of hot topics exhibited a fluctuating downward trend; however, significant national energy policies contributed to this increase. The sentiment tendencies of the themes were relatively positive and showed similar evolutionary patterns, with evident reactions to the acts of energy corporations. Based on the results of this study, suggestions are proposed for energy corporations to actively respond to online public opinion from the perspectives of content relevance, popularity enhancement, and sentiment guidance. The study demonstrates the value of integrated NLP and machine learning methods in analyzing large-scale opinion dynamics and offers actionable insights for corporate communication strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Clustering and multidimensional evolution analysis of hot topics in online public opinion about energy corporations—a case study of Sina Weibo

  • 期刊论文

  • 发表日期：

    2026年6月1日

查看全文
  • Liu Chun,
  • Haoran Mao,
  • Jiakun Wang,
  • Jiaen Shangguan

摘要

In the context of digital transformation in public discourse, computational analysis of evolving hot topics in online public opinion regarding energy corporations provides critical insights for market intelligence and public opinion guidance. This study proposes an innovative framework that integrates topic modeling and multidimensional analytics to decode public perception patterns of energy corporations. Original text data related to eight renowned energy corporations were collected on the Sina Weibo platform from October 1, 2021, to September 30, 2023. By employing the Latent Dirichlet Allocation (LDA) probabilistic topic model, topic clustering was conducted, followed by a secondary dimensionality reduction of clustering results through the integration of inter-topic distances and text analysis. Subsequently, a comprehensive analysis of hot topics (themes) concerning energy corporations was conducted using co-occurrence vocabulary, a heat index, and sentiment analysis with a visual presentation of evolutionary trends. Online public opinion on energy corporations can be categorized into seven hot topics, which can be further summarized into three major themes. The content evolution of hot topics remains relatively stable, with topic generation being driven by leading corporations in new energy businesses. The popularity evolution of hot topics exhibited a fluctuating downward trend; however, significant national energy policies contributed to this increase. The sentiment tendencies of the themes were relatively positive and showed similar evolutionary patterns, with evident reactions to the acts of energy corporations. Based on the results of this study, suggestions are proposed for energy corporations to actively respond to online public opinion from the perspectives of content relevance, popularity enhancement, and sentiment guidance. The study demonstrates the value of integrated NLP and machine learning methods in analyzing large-scale opinion dynamics and offers actionable insights for corporate communication strategies.