Research On Online Coal Mine Accident Case Classification Method Based On Text Mining

The rapid development of information technology has prompted coal mining enterprises to accumulate a large amount of coal mine data resources.Reading the relevant literature on coal mine data mining,we can find that the current data mining objects in coal mines mainly focus on hidden danger data and monitoring data,while the research results of online coal mine accident cases are less,resulting in waste of data resources.The online coal mine accident case is a kind of unstructured data summarized in many aspects such as the time and cause of the accident,and it has certain difficulty in the process of processing.However,a large amount of information contained in the data is of great significance for coal mine safety production and safety management.Therefore,this thesis selects the case of online coal mine accident as the research object,constructs an automatic classification model of coal mine accident case,and mines the information contained in the coal mine accident case.In order to study the automatic classification method of online coal mine accident cases,this thesis studies relevant theoretical techniques.The common text representation model is a statistical language-based word bag model and a Tf-idf model representation method.Word2-vec is a text representation method based on neural network language,which can convert words in Chinese documents into word vectors.In order to construct an automatic classification model,this paper improves Word2vec,converts the output word vector into a document vector,and realizes the conversion of Chinese text data to computer language.The classification effect of SVM classifier is affected by the parameters.In order to construct the classification model by using the optimal model parameters,this paper combines the grid method with the support vector machine to construct the cgSVM classification model,and uses cgSVM to realize the automatic classification of online coal mine accident cases.In order to study the automatic classification method of online coal mine accident cases,three text representation methods are combined with different classifiers to build word-SVM,word-Mul-NB,word-DTC,Tf-idf-SVM,Tf-idf-MulNB,Tf-idf-DTC,Word2vec-SVM,Word2vec-cgSVM 8 text automatic classification models.Compare the contents related to the coal mine accident case in the coal mine monitoring network and the coal mine safety net website,and crawl the coal mine accident case corresponding to the coal mine safety net as the experimental data to verify the model performance of the automatic classification model.The classification effects of the eight models were evaluated from the predicted and actual values of the comparison model and the performance evaluation index values.It is concluded that the prediction result of the Word2vec-cgSVM model is closer to the actual value,and the corresponding accuracy rate,recall rate,and fl-score are 0.977,0.976,and 0.976,respectively.Through research,this paper constructs an automatic classification model for coal mine accident cases.After the model performance analysis,the accuracy ofWord2vec-cgSVM for online coal mine accident cases can reach 97.7%.Applying the Word2vec-cgSVM automatic classification model to the online coal mine accident case classification can save the manpower and time of classification,and has practical significance for improving the classification efficiency of enterprises.The Word2vec and support vector machines are improved,and the improved Word2vec and cgSVM models are proposed.The text representation and text classification model are enriched,which has theoretical significance for the subsequent Chinese text classification research.Figure 14 Table 17 Reference 67...