Text Mining Of Geophysical Exploration Accident Cases And Its Application In Accident Analysis

At present,domestic demand for oil is increasing.Geophysical exploration as the primary link in oil exploration.How to ensure its safe and stable operation has gradually attracted more attention.However,its have strong mobility,more participants and complex operating environment,Causing geophysical accidents to occur frequently.With the development of geophysical exploration projects,a large number of text-type accident cases have been accumulated.How to extract high-value information and apply it to daily safety management is of great significance to reduce the occurrence of geophysical accidents.The language description of the geophysical accident case data is professional and needs to be combined with specific background knowledge to carry out data mining,which is difficult to deal with by conventional structured text methods.In this paper,text mining is carried out with the cases of geophysical accident.The types of accidents,the influencing factors and related relationships of geophysical exploration are obtained,and the law of geophysical accidents is analyzed.The specific research work is as follows:(1)Pre-identification of factors affecting geophysical accidents.The structural composition of the geophysical accident case is analyzed and the dirty data is cleaned,The target of text mining is determined;in order to facilitate the automatic extraction of the late accident influencing factors,Scenario analysis method is used to analyze the occurrence process of geophysical exploration accidents and combined with expert screening and on-site investigation to identify 31 factors affecting the occurrence of geophysical accidents.(2)Geophysical operation accident type clustering identification.According to the text description law of geophysical accident cases,a method based on text clustering for the identification of accident types in geophysical operations is proposed.Firstly,the geophysical accident case is preprocessed;then MIG is used to select the preprocessing results and construct VSM.Finally,Finally,the K-means algorithm is used to cluster the geophysical accident VSM,and the clustering results are manually analyzed to identify the type of geophysical accident.(3)Automatic extraction of influencing factors of geophysical accidents based on semantic framework.Firstly,the accident ontology dictionary of geophysical field is constructed and the basic form of the semantic framework of the influencing factors of geophysical accidents is defined.Then,according to the description characteristics of accidents in accident cases,the semantic framework of 31 accidents is constructed.Finally,the accident ontology dictionary is used to extracted core content from the accident case,and the semantic content matching method proposed in this paper is used to match the core content with the semantic framework of the accident impact factor to realize the automatic extraction of the influencing factors.(4)Analysis of geophysical operations accidents based on association rules.In order to explore the relationship between the influencing factors of geophysical accidents and the types of accidents,the multi-dimensional Apriori algorithm is used to mine the association rules of geophysical accidents.Case analysis was carried out with four types of accidents:traffic accidents,explosions,object strikes,and mechanical injuries.