TIME SERIES FORECASTING METHODS FOR SOCIO-ECONOMIC INDICATORS: A CASE STUDY OF KAZAKHSTAN

Loading...
Thumbnail Image

Date

2024-04-15

Authors

Goloburda, Maiya

Journal Title

Journal ISSN

Volume Title

Publisher

Nazarbayev University School of Sciences and Humanities

Abstract

This study compares traditional statistical methods (ARIMA, ETS) with LSTM, a deep learning approach, to forecast key socio-economic indicators (GDP, Population Growth, Price Index, Income per Capita, Housing prices) in Kazakhstan. Using historical data from the Bureau of National Statistics, the models are trained and evaluated using metrics MAE, MAPE and RMSPE. The research aims to understand the strengths and limitations of each method in the context of Kazakhstan's socio-economic data, providing insights for future forecasting in the region.

Description

Keywords

Type of access: Restricted, Time Series Forecasting, ARIMA, ETS, LSTM, Socio-Economic Indicators

Citation

Goloburda, M. (2024). TIME SERIES FORECASTING METHODS FOR SOCIO-ECONOMIC INDICATORS: A CASE STUDY OF KAZAKHSTAN. Nazarbayev University School of Sciences and Humanities