Research On Visualization Of Miners’ Unsafe Behavior Knowledge Based On Text Mining

The rapid development and wide application of information technology have made the automation construction requirements of the coal industry higher and higher,and the daily production,operation and management tasks of coal mining enterprises have become more and more complex,posing new challenges to coal mine safety production.At the same time,coal mines’ ability to perceive safety information has been greatly improved,accumulating massive amounts of unmined and unanalyzed safety data.At present,the use of collected and stored security data by most coal mining enterprises is limited to simple statistical analysis,and they have not deeply digged and utilized the information and knowledge hidden in the data.In this paper,combined with the data sources of miners’ unsafe behavior and the actual production situation of the coal industry,a research on the visualization of miners’ unsafe behavior knowledge based on text mining is carried out.The main research contents are as follows:First,by analyzing miners’ unsafe behavior management issues and data sources,the type,shape and characteristics of miners’ unsafe behavior data are clarified.Combined with the demand for visualization of miners’ unsafe behavior knowledge,according to the knowledge acquisition process and text mining process,the unsafe behavior of miners is analyzed.The realization process of knowledge visualization of safe behavior is decomposed.Starting from the transformation process of "data-information-knowledge-rule-visualization",a process model of the realization process of knowledge visualization of unsafe behavior of miners based on text mining is constructed.Then,according to the above-mentioned visualization realization model of miners’ unsafe behavior knowledge,the feature extraction,topic mining and spatiotemporal laws of miners’unsafe behavior are studied.Through literature research,expert interviews and semi-structured interviews,the characteristic words of unsafe behavior of miners to be mined are proposed.And the characteristics of unsafe behavior of miners are extracted from coal mine big data by using Python algorithm.Manage the miners’ behavior records manually entered in the system,use the Python word segmentation tool to segment the miners’ behavior data,determine the optimal number of topics for miners’ unsafe behaviors through the cluster analysis results,and use the linear discriminant analysis algorithm(LDA)to complete the miners Unsafe Behavior Topic Mining.Finally,the method of word vector similarity detection and correspondence analysis is used to explore the distribution law of miners’ unsafe behavior in time and space.According to different spatial and temporal granularity and miner’s unsafe behavior data characteristics,an intelligent interactive real-time prediction of miners’ unsafe behavior is constructed.The system realizes the visualization application of miners’ unsafe behavior knowledge.This dissertation uses text mining,data mining and visualization techniques to deeply mine and analyze the information,knowledge and laws implied in miners’ behavior-related data,and make them transparent,intuitive,and visualized.New methods and new approaches to problems such as the expansion of safety behavior data,low information processing capabilities,and insufficient knowledge extraction efficiency.The visualization of miners’unsafe behavior knowledge is an indispensable and important content to promote the visualization of coal mine safety management,which in turn plays an important role in improving the management decision-making power of coal mine personnel,the insight into potential safety hazards and the ability to optimize resources.