A Study on Automatic Indexing Based on Textual Mentions to Geographical Location in Story Archiving

2020 Fiscal Year Final Research Report

• Project/Area Number: 18K11982
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 90020:Library and information science, humanistic and social informatics-related
• Research Institution: University of Tsukuba
• Principal Investigator: INUI Takashi 筑波大学, システム情報系, 准教授 (60397031)
• Project Period (FY): 2018-04-01 – 2021-03-31
• Keywords: 文書ジオロケーション / エンティティ・リンキング / 地名抽出 / 固有表現抽出 / Toponym resolution

Outline of Final Research Achievements

This research project aims to develop a document retrieval technology by geographic location by indexing geographic locations mentioned in the document contents. The main research results are as follows. (1) We developed a deep learning-based geographic name extraction model that is especially robust to unknown words by using word information in documents and image information corresponding to words. (2) We developed a model for identifying the real-world geographic location of place names (place name disambiguation) based on word distributions with data expansion focusing on address hierarchy. (3) By integrating the above models, we have developed a technology to automatically identify geographic location information required for indexing geographic locations mentioned in document contents with a certain level of performance.

Free Research Field

自然言語処理

Academic Significance and Societal Importance of the Research Achievements

本研究課題は、大規模自然災害アーカイブにおいて、従来技術では地理情報システムとの親和性の低かった文書コンテンツに対して、特定の被災地域に限定したコンテンツ検索を実現するための技術開発を目的としたものである。本研究課題で得られた成果を活用することにより、自然災害に対する防災・減災対策や、自然災害からの復旧・復興事業に資する情報へのアクセス効率が従来よりも向上することが期待される。