Template-type: ReDIF-Article 1.0
Author-Name: Natalia S. Nikitina
Author-Workplace-Name: Russian Presidential Academy of National Economy and Public Administration
Title: Forecasting the Real Estate Price Index in Russia
Title: Прогнозирование индекса цен на недвижимость в России
Abstract: This article is devoted to choosing the best model for short-term forecasting of Russia’s real estate price index. Popular machine learning methods: Ridge and Lasso regressions, Elastic Net regression and methods of working with time series were considered: Naive, Exponential smoothing, ARIMA, OLS. The set of variables includes the values of GDP, inflation, effective exchange rate, interbank lending rates, and oil prices. Machine learning methods – Ridge Regression and Elastic Net regression – show the high quality of forecasting the real estate price index compared to standard ways of working with time series – Naive, Exponential smoothing, ARIMA.
The article was prepared in the framework of execution of state order by RANEPA.
Keywords: forecasting, real estate price index, machine learning
Classification-JEL: C32, C53, R30
Journal: Russian Economic Developments
Year: 2022
Issue: 6
Month: June
Pages: 23-28
File-URL: http://www.iep.ru/files/RePEc/gai/recdev/r2250.pdf
File-Format: Application/pdf
Handle: RePEc:gai:recdev:r2250