Research and Application of Key Information Extraction of Speed Restriction Commands for Construction Based on Named Entity Recognition
Dispatching command stands out as the core of railway transport dispatching command.It is a vital means of restriction for train operation,operation mode,and safety precautions.The content of speed restriction commands for construction is in pure text or semi-structured form,resulting in a heavy workload of manual analysis and low efficiency.Information integration and content analysis are inadequate for commands.BERT,BiLSTM,and CRF models were created to conduct named entity recognition of the unstructured content of speed restriction commands for construction and extract key information on the command content to implement information integration and analysis of commands.The models were tested based on the"speed restriction command for the adjacent line"in speed restriction commands for construction.The content of the speed restriction command was analyzed on the basis of the entity recognition results.In addition,the mileage range of train speed restriction and the number of affected trains were calculated.In this way,the extent of the impact of construction on the adjacent line can be efficiently analyzed.The application results indicate that named entity recognition based on BERT-BiLSTM-CRF models provides powerful data support for command content analysis,thus assisting in command and decision-making for railway transport dispatching management.
dispatching commandspeed restriction commands for constructionnamed entity recognitionBERT-BiLSTM-CRF modelskey information extraction
NETL
NSTL
万方数据
