본문 바로가기
  • Home

Development of a Model for Identifying Drug Organizations and Their Scale through Tweet Clustering

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2024, 29(10), pp.207-218
  • DOI : 10.9708/jksci.2024.29.10.207
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : September 9, 2024
  • Accepted : September 30, 2024
  • Published : October 31, 2024

Jin-Gyeong Kim 1 Eun-Young Park 1 Da–Sol Kim 1 Cho-Won Kim 1 Jiyeon Kim 1

1대구대학교

Accredited

ABSTRACT

In this paper, we propose a model for identifying drug trafficking organizations and assessing their scale by collecting drug promotional tweets from the social media platform ‘X,’ with a focus on investigating drug crimes that frequently occur among teenagers and young adults. Recently, various cyber crimes, such as drug distribution, illegal gambling, and sex offense, have been on the rise, exploiting the anonymity provided by social media. Drug trafficking organizations, in particular, operate in a decentralized cell structure, where each member receives anonymous instructions regarding only their specific role and is not directly connected to other members. To track these types of crimes, we designed experimental scenarios using various clustering algorithms, such as K-means Clustering and Spectral Clustering, alongside text embedding models like BERT (Bidirectional Encoder Representations from Transformers) and GloVe (Global Vectors for Word Representation). Furthermore, the clustering results derived from each scenario are validated using Jaccard Similarity and a full-scale investigation. We then analyze tweet clusters identified as the same drug organization across all scenarios, prioritizing the identification of high-priority accounts for cyber investigations.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.