Forecasting of Tourism Demand in Central and Eastern Europe: ARIMAX with Fourier Seasonality and Deep Learning Benchmarks
Serdar CELIK
https://doi.org/10.18267/pr.2026.vol.2587.4
Abstract: This study forecasts monthly international tourist arrivals in ten Central and Eastern European (CEE) countries using a parsimonious time-series framework designed for strong seasonality and pandemic-induced disruptions. Using data from 2004–2025, we estimate an ARIMAX model incorporating a deterministic trend, Fourier seasonal terms, and a COVID-19 intervention dummy. Out-of-sample evaluation for 2024–2025 shows strong predictive performance across countries. Comparative analysis with a BiLSTM–Transformer benchmark indicates that the statistical model remains highly competitive in environments dominated by stable seasonal patterns. Based on the preferred specification, we generate 12-month-ahead forecasts for 2026 that preserve the expected summer peak and provide actionable guidance for capacity planning and sustainable tourism management in the region.
Keywords: machine learning, tourism demand forecasting, tourist arrivals, sustainable tourism.
JEL Classification codes: C22, C53, Z32
Fulltext: PDF
Published by: Prague University of Economics and Business, Oeconomica Publishing House
Year of publication: 2026
Online publication date: 20 May 2026
Copyright: Authors of the papers
ISBN 978-80-245-2587-7
ISSN 2453-6113
Pages 53-65