Risk Evaluation Of Coal-mine Safety Management Based On SEM-FSVM

Mining industry is high-risk. Although recent years China has enhanced coal mine safety management, causing accidents in coal mine to decrease as a whole, which still remains unsatisfactory. For the topics like coal mine safety risk identification, risk evaluation, scholars home and abroad have conducted relevant studies in quantity, and they have achieved a lot. However, just because of a large majority of accidents resulting from ill management, risk factors latent in safety management have to be identified, evaluated sufficiently so that essentialized, delicacy management in coal mine safety management can be enforced indeed. On the other hand, complexity and time-variability in coal mine safety management made application of traditional risk identification, evaluation methods to safety management circumscribed in a sense, especially caused overcoming the subjectivity in safety management and implementing dynamic evaluation insufficientBased on analyzing the deep causes of accident happening frequently in coal mine, the paper discovered risk factors in coal mine safety management widely, through document research, systematic investigation, expert consultation and questionnaire survey identified risk factors latent in this management. Refining the latent risk factors with the help of factor analysis, principal component analysis, the paper built risk evaluation index system of coal mine safety management, established risk path relationship model, and applied Structural Equation Model(SEM)to verification of the model, calculated risk evaluation index weight through SEM path coefficient. Then, combined with coal mine safety management features like nonlinearity, dynamic time variability and small sample, the paper conducted intensive study of SEM, Fuzzy Support Vector Machine(FSVM), and SEM-FSVM evaluation system. The empirical research shows SEM-FSVM evaluation system integrates the advantages of SEM, FSVM, satisfying coal mine safety management rick evaluation better. The study in the paper is expressed as follows.1. Coal mine safety management rick factor identification.After analysis and summary of related literature home and abroad, in combination with coal mine safety management status in China, and through questionnaire survey, expert consultation, and field research, the latent risk factors in coal mine safety management were identified in the study; algorithm Page Rank based on random walk was used creatively to calculate expert weight, which ensuring scientificity and practicality of data.2.evaluation index system establishment and index weight calculation.Reviewing traditional evaluation index weight calculation, the paper stressed defects of existing methods. By virtue of factor analysis and principal component analysis, the paper extracted major risk factors effecting coal mine safety management, established risk evaluation index system in coal mine safety management. Besides, SEM was used to verify risk path relationship model built, SEM path coefficient was analyzed, and index weight was decided accordingly.3. Study of coal mine safety management risk evaluation.Combined with coal mine safety management features like nonlinearity, dynamic time variability and small sample, the paper lucubrated the three risk evaluations SEM、 FSVM and SEM-FSVM, whose algorithm principles, programs and applications to coal mine safety management risk evaluation were analyzed in the paper. Creative combination of FSVM with SEM facilitated establishment of the integration evaluation model of SEM-FSVM. Inner product of both index path coefficient of SEM and kernel function of FSVM was performed (GAUSS is used in the paper.), thus forming index weight kernel function GAUSS, and risk evaluation model SEM-FSVM based on dual weight of index and sample was built. Therefore, the effect of evaluation index contribution and sample contribution on evaluation results was finally removed, guaranteeing scientificity, accuracy of risk evaluation.4. Empirical study.A typical sample coal mine was selected, and empirical study of the coal mine was also made. Contrast results of couples of evaluations reveals that risk evaluation model SEM-FSVM integrated advantages of SEM and FSVM, which both reflects relations among risk factors and preferably highlights the strength of the model in terms of coal mine safety management evaluation. As a result, this model is of higher evaluation accuracy, better satisfying coal mine safety management risk evaluation.5. Major research findings in the paper were concluded finally, and further research orientation to follow was also discussed.The paper is of a distinct feature of connecting theoretical research and practical application, explores and tries evaluation and its application of incorporating SEM with FSVM, and enriches and widen coal mine safety management risk research field. Therefore, the study in the paper is of theoretical significance, application prospect.