Coal Mine Safety Event Ontology And Its Application Research In The Query Expansion

China is the world’s major coal producing country, but also the occurrence of coal mine disasters and up to one million tons mortality in most countries, from the statistical data of coal mine accidents in coal mine accidents still can see Yan Jun, and coal mine accidents result in loss of life, huge economic losses, emergency disposal of its existing resources to share, the effective countermeasures difficult problems, so the emergency management research is very urgent. Coal mine accident has the characteristics of contingency, so difficult to forecast, it is difficult to use a single method and model to solve the emergency decision-making, but the need for the immediate context integrated information, knowledge and model. Coal mine a large number of security incidents(accidents) online reports, from the large number of reports and documents quickly and accurately find the useful information, its difficulty and accuracy is a big problem. Therefore, how to build the model from different disciplines, unstructured or different structure information, knowledge, model of organization and management, effective and easy to establish and maintain the information, knowledge, and according to the context model, provide rapid information, knowledge and model accuracy is one of the key problems of emergency management in coal mine accident.Ontology as a representation of knowledge, can well reveal the concepts in the domain and relations, is a static concept. As human knowledge unit, reflects the things to move, has drawn the attention of experts in many fields of academia. Methods used to describe the static concept is difficult to describe the dynamic events, therefore, the research of event ontology is attracting more and more attention of scholars. Compared with the traditional ontology, event ontology with higher granularity event as a knowledge representation unit, more practical, more convenient for representing and processing knowledge. The event ontology is introduced into the field of coal mine safety events around the event ontology model, ontology building, event oriented query expansion, studies some problems of their existence, the research achievements are as follows:(1) construction of event ontology model. Define the event, the event class, and analyzes the relationship between the six elements and events. On the basis of the various elements of the event class, the relationship between event classes conducted in-depth analysis, reveals the relationship between the existence of an event. To determine the occurrence of an event class instance impact on other event instance of class size factor for the event class. In the definition of the basis between events, event class, event class relations and the influence factors on the event class, event ontology model construction.(2) presents an event recognition method based on multi feature fusion. Event recognition is the basis of event extraction, this paper defines the process of event recognition task, that model has carried on the elaboration to the classifier model and text. In order to improve the recognition effect, put forward a kind of event recognition method based on multi feature fusion, the experimental results are analyzed.(3) construction of coal mine safety events. Study on construction method of domain ontology from a subset of events in the field of emergency. Analysis of dynamic and static conceptual difference between events in the mutual association, according to the event with the time changing characteristic, put forward considering the events of the calculation of Authorities value and Hubs value of the event(denoted as HARank, and the computational identification, this algorithm is applied to text in the collection of important events. The event class importance event ontology.(4) extension method is proposed for event event ontology based query. Coal mine safety event ontology storage of coal mine security event class, event class, event each static attribute relationship between classes, and by expanding the event instance, the formation of a domain ontology of events to the event class prior knowledge, information query can be carried out for the event query expansion. This paper focuses on the recognition of the event, the query term is defined and the discrimination of events, extended event selection, query term weights and query and document similarity calculation, puts forward the query content dependence in events and a determination method based on limited. Text feature selection and weight settings using the vector space model, the event class, event to event elements from the two aspects of this event class spread.