Modelling demographic processes in St. Petersburg
Loading...
Date
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Проведен анализ временного ряда, составленного из значений показателя численности населения города Санкт-Петербурга за период 1970 г. по 2016 г. Построены пять полиномиальных моделей, описывающих динамику показателя численности за период с 1970 г. по 2013 г.. По коэффициенту детерминации определены три модели, наилучшим образом аппроксимирующие данные. Для этих моделей проведен анализ остатков по графикам, значениям среднего и дисперсии. Выбор лучшей модели для прогнозирования осуществлен по критериям Акайке и Шварца. Рассчитана точность аппроксимации модели за период с 1970 г. по 2013 г. и точность модели на проверочной выборке за период с 2014 г. по 2016 г. Построен прогноз численности населения г. Санкт-Петербург на январь 2017 г..
The analysis of time series, composed of index population values in St.Petersburg for the period from 1970 to 2016. Сonstructed five polynomial models, describing the dynamics of the index population values for the period from 1970 to 2013. At the coefficient of determination identified three models, that best approximate the data. For these models, the analysis of residues from the graphs, the values of the mean and variance.Choosing the best model for forecasting carried out by Akaike and Schwarz criteria. Calculated accuracy of the approximation model for the period from 1970 to 2013 and the model accuracy on the test sample for the period from 2014 to 2016,and the model accuracy on the test sample for the period from 2014 to 2016. Completed forecast of population in St.Petersburg on January 2017.
The analysis of time series, composed of index population values in St.Petersburg for the period from 1970 to 2016. Сonstructed five polynomial models, describing the dynamics of the index population values for the period from 1970 to 2013. At the coefficient of determination identified three models, that best approximate the data. For these models, the analysis of residues from the graphs, the values of the mean and variance.Choosing the best model for forecasting carried out by Akaike and Schwarz criteria. Calculated accuracy of the approximation model for the period from 1970 to 2013 and the model accuracy on the test sample for the period from 2014 to 2016,and the model accuracy on the test sample for the period from 2014 to 2016. Completed forecast of population in St.Petersburg on January 2017.