Research On Automatic Construction Of Coal Mine Accident Ontology Based On Text Case

As China's basic energy source,coal often has greater risks in the mining process.Coal mine accidents occur from time to time,and the safety situation of coal mines is becoming increasingly severe.Now in this era of rapid development of information technology,when a coal mine accident occurs,a lot of information about the accident appears.These information contains a lot of useful knowledge that can help to ensure the safety of coal mine production.However,these unprocessed knowledge are disorganized and not conducive to the integration and sharing of knowledge.As a result,it cannot be used well by coal mine safety production management workers,causing waste of information resources.Ontology as a standard and unified way of knowledge expression and description,is suitable for sharing and reuse of knowledge.Although there have been some researches on coal mine safety work based on ontology in the field of coal mine,these ontology are generally constructed manually.Ontology concepts and relationships between concepts are determined manually,and the quality of the constructed ontology is not high and has limitations.In addition,in the face of the explosion of information today,the manual construction method requires a lot of manpower and material resources,and can no longer meet the current needs.Based on the research and analysis of the above problems,this paper uses the coal mine accident case on the coal safety net as the data source,and constructs the coal mine accident ontology by an automatic construction method.The focus is on the use of automatic extraction to obtain the concept of coal mine accident ontology and the relationship between concepts from coal mine accident cases.The main contents of the research are as follows:(1)Through natural language processing technology,coal mine accident text cases are transformed into words,which constitute the candidate concept set of coal mine accident ontology.Using multiple features to describe the concepts in the candidate concept set,a BP neural network is constructed and the feature vector composed of these feature values is used as the input of the neural network.The neural network used to distinguish whether it belongs to a coal mine accident concept is obtained through training,and the neural network is used to screen the concepts in the candidate concept set to obtain the coal mine accident ontology concept.(2)A word vector space model of coal mine accident concepts is constructed based on coal mine accident text cases,and hierarchical clustering based on similarity between concepts is performed.The hierarchical relationship between the concepts of ontology is obtained through the clustering results.For non-hierarchical relationships between concepts,traditional association rule-based methods can only mine conceptual pairs that have relationships,but not specific relationships.Therefore,this paper adopts the method of extended association rules to filter out verbs that meet the conditions between concepts with association relations based on association rules.The verb is used as the relationship label between concept pairs,so as to obtain the non-hierarchical relationship between the ontology concepts.(3)After the concept of coal mine accident ontology is obtained through automatic extraction,the coal mine accident ontology can be obtained.Before applying the ontology,you need to verify whether the constructed ontology meets the application conditions.The composite standard evaluation method is used to evaluate the coal mine accident ontology.The evaluation results show that the constructed ontology can meet the requirements well,and it also verifies the effectiveness of the automatic construction method in this paper.Form the formal description of the coal mine accident ontology by OWL ontology language and store it as an OWL format text file to facilitate the sharing and reuse of the ontology.Finally,the OWL text file of the coal mine accident ontology was imported into Protege to realize the visualization of the ontology,and based on the visualization,the application of knowledge query of coal mine accident based on ontology was realized.This paper makes the proposed automatic construction method of coal mine accident not only save a lot of manpower and material resources,but also automatically extracts the concept of ontology and the relationship between concepts from actual coal mine accident cases,which can make the constructed ontology more comprehensive and closer to actual.The effect will be better in practical applications later,and it can provide effective help for coal mine accident management and coal mine safety production management.Figure 22 Table 15 Reference 71...