关键词: 煤矿安全隐患, ERNIE-BiGRU-CRF算法模型, 命名实体识别, 信息抽取

Abstract: In order to fully explore the key text knowledge of coal mine safety hidden danger and help the safety management personnel of coal mine enterprises to better carry out hidden danger investigation and management work, a named entity recognition method based on pre-training language model is proposed. Firstly, entity categories of coal mine safety hidden danger were defined according to the new version of Coal Mine Safety Regulations and Criteria for Determining Potential Major Accidents in Coal Mines, and 7 entity categories and 15 entity labels were constructed using BIO labeling strategy. Then, the collected data of coal mine hidden danger investigation are preprocessed, and relevant entities are manually marked by experts in the field of coal mine safety, and 1500 standard data sets of named entities for coal mine safety hidden danger are obtained. Finally, the ERNIE pre-training model is used to represent the text word vector of coal mine safety hidden danger, and the BiGRU structure is used to extract the context semantic features and the CRF model is used to decode the entity label, and complete the research on the named entity recognition of coal mine safety hidden danger. The experimental results show that: The accuracy rate, recall rate and F1 value of ERNIE-BiGRU-CRF model on sequence labeling tasks are 56.69%, 69.23% and 62.34%, which are respectively 6.85%, 13.74% and 9.83% higher than baseline model of BiLSTM-CRF. And there is little difference between the entity extraction result and the actual label result. In addition, the ablation experiment also verified that BiGRU layer could better capture semantic dependencies of text context for coal mine safety hidden danger and CRF layer could further optimize label sequence. It can be seen that the named entity recognition model based on the ERNIE-BiGRU-CRF algorithm structure has a good entity recognition result in the text information extraction of coal mine safety hidden danger, which provides convenience for the accomplishment of intelligent management of coal mine safety hidden danger.