Core Time

[P-16] Study on Emergency Management Using Incident Response Log Classification Based on Activity Objectives

*Naoko Kosaka¹, Takeshi Yamaguchi², Tomohiro Kokogawa¹, Satoshi Kubota¹, Kentaro Inui² (1. NTT, 2. Tohoku university)

キーワード：EOC(Emergency Operations Center), Emergency Management, Natural Language Processing

It is necessary for an EOC (Emergency Operations Center) to observe the whole the damage from a disaster, rather than a detailed situation of the site, and assess the impact on the organization. Therefore, EOC staff should determine the overall situation of the organization from the reports of individual organizations and allocate appropriate resources to appropriate locations.

FEMA in U.S.A. has established the ESF (Emergency Support Functions)[1] which standardizes the emergency support function in the incident response. By ESF, it becomes possible to connect to quick observation of the situation and efficient and effective resource allocation, because the incident response between organizations is unified in the axis of ESF.

In this paper, we propose a method to observe situations using incident response log on the “KADAN,” emergency management support system[2][3], based on activity objectives as an ESF-like axis, applying natural language processing.

Referrences

[1] FEMA, “Emergency Support Functions (ESF), ” https://www.fema.gov/media-library/resources-documents/collections/533

[2] Naoko Kosaka et. al., “Disaster Information System Using Natural Language Processing,” JDR, Vol.12, No.1, pp.67-78, 2017.

[3] Naoko Kosaka et. al., Applicability Assessment of An Emergency Management Support System “KADAN,”DRRSC2019.