Abstract: In order to explore coal mine safety management model from a perspective of smart mines, a PSR model of coal mine safety management was established by introducing concept of PSR from the field of ecological safety in it, and relationship among risk factors, coal mine safety status and safety risk management & control was revealed. Then, multi-source heterogeneous data collected by coal mines was analyzed, and TF-IDF hidden danger classification model and PCA coal mine safety assessment model were constructed based on framework of the PSR model. Finally, this model was applied in a mine in Inner Mongolia to verify its effectiveness. The results show that TF-IDF algorithm-based hidden danger classification model and PCA safety assessment model can comprehensively analyze safety management effect of coal mines, and main hidden dangers can be checked according to analysis results.

Key words: multi-source heterogeneous data, coal mine safety management, effect evaluation, pressure-state-response(PSR) model, term frequency-inverse document frequency(TF-IDF), principal components analysis(PCA)