Research on text analysis of hidden dangers of coal mine accidents based on link prediction
In order to effectively mine the valuable coal mine accident hidden danger information from massive text data and conduct prediction,the pre-processing text of safety hidden danger generated in the production process of a certain coal mine was studied. The link prediction method was used to explore the association rules of coal mine accident hidden danger texts,and a local random walk index combining RA and RWR was proposed. The results show that the improved index has better ac-curacy than the local similarity method within a higher removal ratio range,and its calculation efficiency is significantly better than the global random walk index. By analyzing the co-occurrence relationship of keywords,the accuracy of the improved link prediction method is verified,and some node relationships that exist in future but have not yet been generated at present are successfully predicted. The research results can provide a new method and tool for work safety management in coal mines,which is conducive to improve the work safety level of coal mines.
万方数据
维普
