Title: A hazardous chemical knowledge base construction method based on knowledge graph

Authors: Guanlin Chen; Qiao Hu; Qi Lu; Kaimin Li; Bangjie Zhu

Addresses: School of Computer and Computing Science, Zhejiang University City College, Hangzhou, 310015, China; School of Computer, Zhejiang University, Hangzhou, 310027, China ' School of Computer and Computing Science, Zhejiang University City College, Hangzhou, 310015, China; School of Computer, Zhejiang University, Hangzhou, 310027, China ' China National Air Separation Engineering Co., Ltd, Hangzhou, 310051, China ' City Cloud Technology (China) Co., Ltd, Hangzhou, 310000, China ' School of Computer and Computing Science, Zhejiang University City College, Hangzhou, 310015, China

Abstract: This paper proposes a model for extracting risk information of hazardous chemicals and constructs a hazardous chemicals knowledge graph. This paper constructs a hazardous chemicals risk information dataset. In this paper, the proposed model combines word-feature into character-feature as character's embedding; uses a joint model of BiLSTM and self-attention mechanism to encode characters and a bidirectional label distribution transfer model is used to decode the classification. Using basic data and the risk information extracted by the model, this paper establishes a knowledge graph. The experimental results show that the model has a better effect on the extraction of risk information than classical models, and the knowledge graph is also more comprehensive in the knowledge system. Combined with sensors, the management system based on this knowledge graph can determine whether a chemical in a warehouse is in a safe environment and whether a chemical meets the storage conditions.

Keywords: hazardous chemicals; knowledge graph; NLP; NER; sensor.

DOI: 10.1504/IJRIS.2022.126656

International Journal of Reasoning-based Intelligent Systems, 2022 Vol.14 No.4, pp.184 - 193

Received: 01 Apr 2022
Accepted: 21 Jun 2022
Published online: 31 Oct 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article