An Automatic Constructed Subway Construction Safety Risks Transmission Networks: A Pipeline Model Based on Named Entity Recognition and Relation Extraction

Abstract Safety risk management is a critical part during the subway construction. However, conventional methods for risk identification heavily rely on experience from experts and fail to effectively identify the coupled relationship between risk factors and events embedded in accident texts. Consequently, they are unable to provide substantial guidance for subway safety risk management. The research developed a domain entity recognition model for subway construction safety risks based on Bidirectional Long Short-Term Memory Networks with Conditional Random Fields (BiLSTM-CRF) and a domain entity causal relation extraction model based on Convolutional Neural Networks (CNN), with a dataset consisting of 562 instances of subway construction accidents. Constructed models achieved automatically extract safety risk factors, safety events, and their causal relationships from subway accident texts. The results demonstrated that the precision, recall, and F1 scores of MCSR-NER-Model all exceeding 77%. Its performance in the specialized domain named entity recognition with a limited volume of textual data is satisfactory. The MCSR-CE-Model achieved an impressive accuracy, recall, and F1 score of 98.96%, exhibiting excellent performance. Moreover, the extracted entities are normalized and a domain dictionary was developed. Based on the processed entities and relationships processed by the domain dictionary, 533 domain entity causal relation triplets are obtained, facilitating the establishment of a directed and unweighted complex network for subway construction safety risks and a case database. This research successfully converted subway construction accident texts into a causal chain structure of "safety risk factors to risk events," providing detailed categorization of safety risks and events. Concurrently, it revealed the interrelationships and historical statistical patterns among various safety risk factors and categories of risk events through the complex safety risk network. The establishment of the database furnishes project management personnel with effective response measures for historical risk events.