This paper proposes a basic concept of historical big data structuring: the “mention-entity model,” a data model responsible for the bidirectional linking of document space and entity space, and its database system, Rekuroku. In addition, we focus mainly on the semi-structuring and quantification processes among the historical big data structuring workflows and propose a markup method using “Minna de Annotation” and a seismic intensity determination method using Large Language Models (LLM). Finally, we apply these results to analyze damage records from the Ansei-Edo Earthquake and discuss digital criticism in historical big data research.

      @InProceedings{ jm24b,
      author = {Asanobu KITAMOTO and Yuta HASHIMOTO and Junzo OMURA and Yasuyuki KANO},
      title = {Structuring Historical Big Data and Analysis of Damage Records on the Ansei Edo Earthquake},
      booktitle = {Proceedings of IPSJ SIG Computers and the Humanities Symposium 2024},
      pages = {37-44},
      year = 2024,
      month = 12,
      doi = {},
      note = { (in Japanese)},
      }