Construction of Knowledge Graph on Debris Flow Prevention Domain

Abstract

Debris flow disaster breaks out frequently and causes serious harm. Therefore, it is of great significance to construct of debris flow prevention knowledge graph to work on disaster prevention and mitigation. Aiming at the problem that the cognitive knowledge correlation in the field of debris flow disasters prevention is not strong, this paper not only proposes a method to construct a knowledge graph of debris flow disasters prevention from the data layer, technology layer, and application layer but also divides debris flow theoretical knowledge, disaster prevention strategies, debris flow disaster events and debris flow method models into four entity types and analysis correlation on the relationship between the four entity types. BiLSTM-CRF method and template matching method are used for knowledge extraction of the above four entities. The experiment result shows that 1233 debris flow entities and 2797 entity relationships are extracted. The accuracy rate of the extracted debris flow entities is about 80%. Finally, the neo4j graph database is used to store the extracted entities and relationships, and a knowledge graph of debris flow prevention for disaster prevention is constructed, which realizes the query and retrieval of debris flow prevention and control knowledge.