Research On Work Safety Accidents Data Analysis Based On Data Mining Technology

The accident investigation system's data of the State Administration of Work Safety show that:a total of 28115 accidents occurred in our country,resulting in 16059 deaths or whereabouts unknown during January to July of 2016.Data mining tools have been widely used in business,while it is rarely used in the field of work safety.With the development of history,work safety database of China accumulated a lot of text,pictures and audio data about the accidents,which can play a significant role in the prevention of safety accidents.However,the data collected is very difficult to explore the depth of the information contained therein because it is too fragmented,thus it is always a major problem plagued the security academia to find useful information efficiently from large data.This paper starts from the data of work safety in China,three things are accomplished including the statistical analysis of the number of accidents and accident deaths,number of accidents prediction,correlation analysis of the location,the type of accidents and the time of occurrence.Theory and method of data mining,such as cluster analysis,association rule mining and time series prediction are used to analyze and excavate the work safety accident data to explore the relationship between safety accidents and various factors..The data is obtained from the website of the State Administration of Work Safety through octopus,which contains text data and digital data.Text data must be pre-processed by the text mining process before statistical analysis and excavation:The statistics of the number of deaths and location information,which can get the access to the basic rules of work safety accident information;then text clustering and correlation analysis are made for the text data to mine the rules and mechanism of accident data in a deep level.Using RHadoop cloud platform to deal with work safety data,and compared the time-consuming with single computer;Finally,in order to improve the accuracy of the accidents number forecast,a time series combinatorial forecasting model based on ARIMA and LOESS is proposed.