Аппроксимация продолжительности жизни в России на основе интервальных данных и минимаксной модели

Abstract

Article provides automated processing of interval data on dynamics life expectancy in Russians over a long period of time (more than 100 years). Software product for approximation of dynamic processes in the demographic sphere that occur under unstable trend conditions is developed. Software product is based at the author's model of spline approximation the interval data by linear polynomials. The algorithm of program uses method for finding the moments of joining linear splines using reasonable properties of solving minimax problem for dynamic series whose values are interval data: the lower limit of the interval is fixed life expectancy of men, the upper limit is fixed life expectancy of women. As result of computational experiments, the jump-type curve approximation obtained. The jump-type curve preserves important properties of modeled series and has good approximation properties. The special feature of developed software product is the ability to get an approximating function in a matter of seconds, which has high accuracy and allows you to analyze profile of demographic situation in Russia, taking into account life expectancy of men and women over a long period. Were identified key turning points (mid 20th century and beginning of 21st century, when there was decline in life expectancy of men and his essential distance from a more stable indicator for women), after which dynamics of indicators have shown an increasing trend of both indices and reducing the gap between them.