The Subgenre Classification of Modern Japanese Novels Using Topic Modeling (LDA) ：An Exploratory Study

HIBI, Yoshitaka

doi:2024.19.1.59

The Journal of Japanese Language Literature Studies > Volume 19(1); 2024 > Article

Special Issue Article

Border Crossings: The Journal of Japanese-Language Literature Studies 2024;19(1): 59-84.
doi: https://doi.org/10.22628/bcjjl.2024.19.1.59

The Subgenre Classification of Modern Japanese Novels Using Topic Modeling (LDA) ：An Exploratory Study

Yoshitaka HIBI

Professor, Nagoya University

トピックモデルを利用した近代日本小説のサブジャンル判定の試み

日比嘉高

名古屋大学大学院人文学研究科教授

Correspondence Yoshitaka HIBI ,Email: yosh.hibi@gmail.com

Published online: 30 December 2024.

This is an open access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( http://creativecommons.org/licenses/by-nc/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

ABSTRACT

This paper explores the application of topic model-based document classification to the study of literary history, focusing on the classification of modern novels into thematic categories that function as “subgenres.” By conducting a topic model analysis on a dataset of full-text digital files of modern novels from the National Diet Library of Japan, I investigate how topics can be interpreted as potential subgenres within the broader literary context. Section 1 reviews traditional perspectives on genre in literary criticism and research, establishing a foundation for identifying and positioning subgenres within the framework of topic modeling. Section 2 builds upon this by defining the criteria and methodologies for determining subgenres through topic models. In Section 3, we conduct a concrete analysis using Latent Dirichlet Allocation (LDA) to examine a selection of full-text files of modern novels from the 1870s to the 1940s, employing a diachronic approach. Section 4 then narrows this diachronic analysis by examining decade-specific trends across the same period. While this study remains in a preliminary phase, Section 5 discusses the potential implications and future directions that topic modeling analysis may offer to literary historical research.

Keywords: genre, Latent Dirichlet Allocation, literary history, modern Japanese novel, topic modeling

キ―ワ―ド: ジャンル, LDA, 小説史, 近代日本小説, トピックモデル