Scholarlink

In this study, we proposed an original keyword extraction method that extracts important keywords for each speaker, applied it to parliamentary session minutes, which contained many statements by politicians, and presented a system to search for familiar politicians based on the content of their statements. In this study, we defined WI-Score as a new important word extraction method. Important words in this study refer to a specific group of words that are more characteristic than those in the statements of other people among the words contained in the statements of the target person. This made it possible to compare and analyze important words from the same perspective as that of the subject. In this study, we applied this method to Japanese politicians and extracted important words from each politician’s speech using data from speeches in the Diet. By visualizing them using principal component analysis (PCA), we realized a system that can detect politicians who speak across parties and faction boundaries. However, this can be applied not only to politics but also to SNS analysis.

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

Discovering Different Parties, Factions, and Close Politicians Using a Word Importance Derivation Method for Parliamentary Proceedings Data

  • 章节

  • 发表日期：

    2025年

查看全文
  • Shota Tamaru,
  • Ryotaro Okada,
  • Ayako Minematsu,
  • Takafumi Nakanishi

摘要

In this study, we proposed an original keyword extraction method that extracts important keywords for each speaker, applied it to parliamentary session minutes, which contained many statements by politicians, and presented a system to search for familiar politicians based on the content of their statements. In this study, we defined WI-Score as a new important word extraction method. Important words in this study refer to a specific group of words that are more characteristic than those in the statements of other people among the words contained in the statements of the target person. This made it possible to compare and analyze important words from the same perspective as that of the subject. In this study, we applied this method to Japanese politicians and extracted important words from each politician’s speech using data from speeches in the Diet. By visualizing them using principal component analysis (PCA), we realized a system that can detect politicians who speak across parties and faction boundaries. However, this can be applied not only to politics but also to SNS analysis.