ПІДХІД ДО ВИЗНАЧЕННЯ РОЗМІРІВ ГОРИЗОНТУ ПРОГНОЗУВАННЯ ДЛЯ ТРЕНДОВИХ МОДЕЛЕЙ

Abstract

The article substantiates an approach to determining the size of the forecasting horizon for making forecasts using trend models. In contrast to the known approaches, the proposed approach takes into account the rate of development of the process under study and the variance of accumulated estimates and is based on controlling the stay of the predictive estimate within a certain interval. The sizes of the interval are determined by the selected quantile of the standard deviation of the values of the random variable. The author substantiated a set of calculated expressions to obtain estimates of the permissible forecasting horizon for trend models. The approach proposed in the article does not apply to constant functions, since they do not reflect the rate of change of processes. It is proposed to carry out the reliability of forecast estimates obtained on the calculated forecasting horizon according to the classical scheme - through the size of the confidence interval calculated for a certain level of significance, which is chosen by the decision-maker. For this purpose, the article provides expressions for calculating the standard mean square error for a linear function, as well as polynomials of the 2nd and 3rd orders. The approach developed in the article is proposed to be used in the practical activities of the management bodies of law enforcement formations during the forecasting of trends in the development of the operational situation in the field of public order protection, as well as ensuring public safety in the country's region.