Construction of Knowledge Graph for Emergency Resources

Heng Mu,Peng Wu,Wenyi Su

DOI: https://doi.org/10.1155/2024/6668559

IF: 8.993

2024-02-12

International Journal of Intelligent Systems

Abstract:Knowledge graphs can effectively organize and represent information related to emergency resources for unforeseen sudden events. In this study, we construct a model layer for the knowledge graph of emergency resources, focused on sudden events, through the classification and analysis of unforeseen disaster measures. This study defines eight interconnected entity types, each characterised by a set of attributes and engaging in one or more relationships with other entity types. Utilizing 121 incident investigation reports from the emergency management departments of various provinces and cities over the past five years, we select five entities with the highest frequency of occurrence along with their corresponding four relationships. We then design an extraction plan for these entities and relationships. Based on the completed knowledge graph data, we formulate 14 questions related to emergency resources for sudden events and construct 19 corresponding question-and-answer templates using a template-based question-answering (QA) approach. We retrieve the corresponding Cypher statement templates through template mapping and obtain the question answers through querying. Finally, we design a knowledge graph question-and-answer system using the Django web framework, which includes entity queries and knowledge QA functions, specifically for emergency resources related to sudden events.

computer science, artificial intelligence

What problem does this paper attempt to address?

The paper aims to address the effective management and rapid response of emergency resources during sudden incidents. Specifically:

1. **Problem Background**: Sudden incidents (such as natural disasters, accidents, public health events, and social security events) continuously threaten the safety and development of human society and the economy. For example, the "9/11" terrorist attacks and the ongoing COVID-19 pandemic have had a significant impact on the world.

2. **Existing Problems**: There are deficiencies in China's emergency resource management, including insufficient investment, low resource integration, low management levels, weak grassroots emergency response capabilities, and insufficient public risk awareness.

3. **Research Objectives**: By constructing an emergency resource knowledge graph for sudden incidents, the aim is to integrate scattered data and achieve efficient information retrieval. This knowledge graph can help government departments improve the accuracy and completeness of emergency resource information and more effectively allocate emergency resources in subsequent incidents.

4. **Specific Methods**: The authors defined eight interrelated entity types and used data from 121 emergency incident investigation reports over the past 5 years to extract entities and relationships. Based on this data, a knowledge question-answering system was designed to provide specific questions and answer templates regarding emergency resources.

In summary, the paper is dedicated to optimizing the management and response mechanisms of emergency resources during sudden incidents by constructing an emergency resource knowledge graph.