Статистические подходы к анализу и прогнозированию демографических данных
Introduction. Possibilities of application ARIMA-models to analysis and forecasting of demographic time series were considered in the article. Foreign studies had shown that the ARIMA-models give good results for forecasting indicators such as population, birth rates and death rates, life expectancy, along with the traditional demographic methods (cohort-component approach). Research technique. Box-Jenkins methodology of the analysis and forecasting of time series, particularly with regard to demographic data: total fertility rate in Russia (1990-2014), the number of marriages by months in Russia (2005-2015), total fertility rate in France (1740-2014) and the unemployment rate in Russia (1996-2016) was used in the work. ARIMA, ARIMA and ARIMA-models, depending on the nature of the dynamics of the studied indicators were analyzed. Results. The analysis had shown that the estimated ARIMA-models for the total fertility rate and number of marriages were adequate and had good statistical and prognostic properties. Forecasts were built on basis of the obtained models. In the case of long series availability of properties with a long memory processes have not been identified.