´ëÇѾð¾îÇÐȸThe Linguistic Association of Korea

ÇÐȸÁö

  • Ȩ
  • ÇÐȸÁö
  • ³í¹®ÀÚ·á½Ç

³í¹®ÀÚ·á½Ç

Á¦¸ñ Analyzing Suicide Notes with Forensic Linguistics and Deep Learning Techniques
ÀúÀÚ Yong-hun Lee & Gihyun Joh
±Ç/È£ Á¦31±Ç / 2È£
Ãâó 101-122
³í¹®°ÔÀçÀÏ 2023-06-30
ÃÊ·Ï Lee, Yong-hun & Gihyun Joh. (2023). Analyzing suicide notes with forensic linguistics and deep learning techniques. The Linguistic Association of Korea Journal, 31(2), 101-122. This paper provides an analysis of suicide notes and ordinary texts using forensic linguistics and deep learning techniques. For the analysis, two types of corpora were compiled. One corpus was composed of suicide notes (SNs), and the other was for ordinary texts (OTs). Seven files were included in the first group, and eight files were contained in the second group. After these two types of corpora were compiled, each text in the corpora was linguistically analyzed with Linguistic Inquiry and Word Count (LIWC). Since the analysis results had 72 dimensions per text, both PCA and t-SNE (dimensionality reduction techniques in deep learning) were applied for the visualization of results. Then, the results were analyzed. Through the analysis, the following facts were observed: (i) suicide notes could be distinguished from ordinary texts, (ii) even though the same author wrote both types of texts, suicide notes could be distinguished from ordinary texts, and (iii) the novels with the 1st person protagonists point of view were also different from the suicide notes, though both types of texts preferred to use the 1st person pronoun I.
ÆÄÀÏ PDFº¸±â  ´Ù¿î·Îµå