Scholarlink

<p>The rapid growth of internet usage worldwide has created an open platform for individuals to express their opinions across various online channels, including Twitter, Facebook, mobile applications, and forums. These social interactions generate vast amounts of data daily. Sentiment analysis, a key technique in natural language processing, involves extracting and analyzing people's opinions, emotions, and reactions to a wide range of subjects such as products, services, events, and public figures. This paper presents a comprehensive survey of sentiment analysis techniques, datasets, applications, tools, and the challenges that shape this evolving field. We provide a detailed exploration of the sentiment analysis process, from data collection and preparation to presenting results through visualization techniques. The paper categorizes sentiment analysis methods into four main types: lexicon-based, machine-learning-based, hybrid, and other approaches, and compares them based on classification performance. Finally, we highlight some of the major challenges in sentiment analysis and suggest potential research directions for the future.</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

Sentiment analysis survey: datasets, techniques, applications, tools, and challenges

  • 期刊论文

  • 发表日期：

    2025年6月21日

查看全文
  • Mustapha Hankar,
  • Toufik Mzili,
  • Mohammed Kasri,
  • Abderrahim Beni-Hssane

摘要

The rapid growth of internet usage worldwide has created an open platform for individuals to express their opinions across various online channels, including Twitter, Facebook, mobile applications, and forums. These social interactions generate vast amounts of data daily. Sentiment analysis, a key technique in natural language processing, involves extracting and analyzing people's opinions, emotions, and reactions to a wide range of subjects such as products, services, events, and public figures. This paper presents a comprehensive survey of sentiment analysis techniques, datasets, applications, tools, and the challenges that shape this evolving field. We provide a detailed exploration of the sentiment analysis process, from data collection and preparation to presenting results through visualization techniques. The paper categorizes sentiment analysis methods into four main types: lexicon-based, machine-learning-based, hybrid, and other approaches, and compares them based on classification performance. Finally, we highlight some of the major challenges in sentiment analysis and suggest potential research directions for the future.