Mathematical modelling and prognozing of safety labor parameters based on real statistic data

The article discusses the identification of occupational exposure using regression analysis and models of neural networks based on ROSSTAT statistics on occupational diseases and work conditions for industry of the Russian Federation «Mining of hard coal, brown coal and peat». When comparing the predictive values of risk indicators derived from these models throughout the original sample data, it follows that if the difference in absolute values, the order of predicted values match. For models described by the authors of the developed software product «Neural network predictor», received the certificate on the State registration. This software module is used in the automated system of safety management officers of the Federal fire service of EMERCOM, establishment and approbation which is scheduled at the Academy of the State fire service of EMERCOM in 20162020 timeframe.