Research On The Application Of NLP In Intelligent Alarm Receiving And Handling System

With the continuous acceleration of urban industrialization and modernization,the expansion of urban scale and the sharp increase in engineering construction have caused frequent accidents.The rescue response speed to the fire rescue team,rescue resource dispatching and allocation,on-site rescue operations,scientific rescue,and plans for key units Comprehensive capabilities such as input put forward higher requirements,and more and more information-based methods are fully integrated with fire service management and actual combat applications.In order to meet the requirements of the city fire command center for receiving and handling police,it is aimed at the efficiency of information input and force deployment of most current alarm receiving systems.Low-level issues.This thesis proposes an ALBERTbased pre-training model to extract the elements of the police information in the alarm receiving system,and then conduct experiments on the fire alarm information corpus built by itself,and construct the ALBERT-Bi LSTM-CRF fire alarm text entity recognition model,design An entity recognition scoring mechanism based on radicals is used to further improve the recognition effect of the fire alarm elements.Its F1 value is 81.660.Because this field needs to quickly extract the voice information after converting it into text,it is verified that the model can improve the efficiency of alarm input and assist decision-making while ensuring the time cost and recognition accuracy at the same time.The goal of this project is to apply natural language processing technology to the fire alarm reception and handling system,improve the efficiency of alarm reception and handling as much as possible,and try to build a knowledge map of fire alarm information elements.Based on this goal,this topic explores the following contents:（1）First,it introduces the current situation of firefighting and handling,and investigates the research status of natural language processing technology in the field of firefighting,and then introduces the significance of applying natural language processing technology to the firefighting and handling system.Then the related theories and techniques are introduced.（2）Regarding how to construct a data set of fire alarm texts,as well as the preprocessing of the data,a better data set can be obtained through data enhancement and other means.（3）By comparing the performance difference between the ALBERT-Bi LSTM-CRF model and the BERT-Bi LSTM-CRF,Bi LSTM-CRF model in named entity recognition,it verifies the fire alarm text entity recognition task based on the ALBERT model It also introduces a radicalbased scoring mechanism to improve the model’s effectiveness in identifying entities with alert elements.（4）Preliminary design and implementation of the fire-fighting element extraction module,through the simulated voice-to-alert text as the input of the system,it was verified that the fire alarm text entity recognition module based on the ALBERT-Bi LSTM-CRF model can recognize the fire alarm information text entity.Circumstance,and mark the location of the police situation on the GIS map,and through the entity identification of fire alarm information elements,the data in the data set is extracted,and the partial knowledge map of fire alarm information elements is preliminarily constructed.