Study On The Theory And Methods Of Accident Statistics Analysis And Prediction

In this dissertation, the classification satandard for the accidents related to work-safety is put forward, and the industry accident analysis statistics indicator system is set up. Sorts of accident data are statistically analyzed historically and it is important for us to recognize such status and help us to make decision scientifically and design target. At last accident-forecasting approach is studied. Brown thrice exponential smoothing forecasting model and fuzzy gray forecasting approach are chosen which are regarded important for accident forecasting.The forecasting results of main sorts of accident type concluded in the paper are shown as follows.1. Brown thrice-exponential smoothing forecasting model is applied to forecast sorts of industry accident death totality indicators in the paper.2. Gray forecasting model is applied to predict analyzing sorts of long-term accident death totality ranging from 2003 year to 2020 year in the paper. The average increasing rate per year is up to about five percentages.3. Gray forecasting model is applied to predict analyzing middle and long-term domestic industry accident death totality ranging from 2003 year to 2020 year. The result shows the future industry enterprise accident death totality has increasing trend, which is less than 2 percentages.4. Gray forecasting model is applied to predict analyzing domestic death rate per hundred thousand persons ranging from 2003 year to 2020 year.5. Gray forecasting theory is applied to the time series forecast of the death per kilo man rate and drill accident number in petroleum industry in the paper. The forecasting result is better adaptive to the reality with less error.