������ ������������� ������������ �����ֲ� ������������ ��������� �� ����² ��������� ������ ��������� ������Ҳ�

Abstract

To predict the occurrence of emergencies of man-made character in the state widely used methods of regression analysis. The regression model of such a process is usually non-linear. It is substantiated that the regression model of the occurrence of emergencies of man-made character should be presented in the form of a power polynomial. At the same time, the degree of polynomial should be about five times less than the number of statistical data.When estimating the parameters of a model using the least squares method, the dispersion of residuals for each observation or group of observations is not always ensured. This leads to the fact that the parameters of the regression model will not have a minimum dispersion, which degrades the forcast accuracy.The article improves the method of determining the predictive value of the number of man-made emergency situations taking into account the forecast of errors in the regression model and refining the estimates of its parameters on the basis of the weighted method of least squares, which allows to improve the accuracy of the forecast.The model of predicting of emergencies of man-made character on the basis of the weighted method of least squares is proposed, which differs in that it takes into account the error forecast of the sample regression model On the basis of the statistical data of the monitoring of emergencies of man-made character in Ukraine, experimental studies have been carried out on the effectiveness of applying the weighted method of least squares to improve the accuracy of the forecast using regression models. As a result of the research it has been established that the method allows increasing the accuracy of the forecast of emergencies of man-made character in the state due to the prediction of model errors and the refinement of its parameters by a weighted method of least squares by 4%As a benchmark for the comparative analysis, the average value of the forecast error module has been selected for the statistical data of the monitoring of emergencies of man-made character in Ukraine during the observation period.