본문 바로가기
  • Home

A Study on Genre Structuring to Improve Unstandardized Classification in Content Platforms

  • Journal of Internet of Things and Convergence
  • Abbr : JKIOTS
  • 2025, 11(4), pp.85~90
  • Publisher : The Korea Internet of Things Society
  • Research Area : Engineering > Computer Science > Internet Information Processing
  • Received : July 4, 2025
  • Accepted : August 5, 2025
  • Published : August 31, 2025

Chae-Jin LIM 1 Jung Soo Han 2 Hyun-Seob Lee 2

1백석대학교 소프트웨어융합전공
2백석대학교

Accredited

ABSTRACT

The content platform utilizes genres and tags as important metadata along with the name of the work and the name of the writer in its content classification. Its purpose is to clearly indicate the basic properties of the content and to combine similar-looking content to increase the accessibility of users. However, these two pieces of information are often randomly assigned by service companies or distributors, revealing problems in structure and consistency. In order to find a way to improve this problem, this study hierarchically clustered the unstandardized genre structure by using the similarity between genres calculated through the PMI (Pointwise Mutual Information), an indicator of statistical correlation between elements, based on the serial information of Naver Webtoon, a representative webtoon platform in Korea from 2005 to 2024. As a result of the experiment, the genre prediction performance through plot embedding in the redefined cluster → genre structure was 91.4%, which was 1.49 times better than when it was not redefined. This study contributes to the improvement of indicators for the classification of unstructured data, the expansion of the general content classification system not limited to webtoons, and the maintenance of the recommendation system.

Citation status

* References for papers published after 2024 are currently being built.