Predicting tourism growth in Saudi Arabia with machine learning models for vision 2030 perspective

Area of Science:

• Data Science
• Econometrics
• Tourism Studies

Background:

• Saudi Arabia's Vision 2030 prioritizes tourism for economic diversification beyond oil.
• Accurate tourist demand forecasting is vital for sustainable development and infrastructure planning.

Purpose of the Study:

• To investigate machine learning (ML) techniques for predicting city-level tourist arrivals in Saudi Arabia.
• To evaluate the performance of various regression and ensemble ML models.

Main Methods:

• Utilized a city-level dataset from 2021-2023.
• Evaluated Random Forest, Gradient Boosting, HistGradientBoosting, and ensemble models (stacking, voting).
• Employed mixed-year training, temporal holdout, and time-series cross-validation.

Main Results:

• The VotingR2 ensemble model demonstrated superior performance across all evaluation scenarios.
• Achieved R² values of 0.9601 (Scenario A), 0.8735 (Scenario B), and 0.9578 (Scenario C).
• Generated long-horizon tourism demand projections for 2024-2034 using Prophet and ML.

Conclusions:

• Ensemble ML methods effectively capture complex, nonlinear patterns in tourism demand.
• Findings offer actionable insights for strategic decision-making, infrastructure optimization, and policy development.
• Supports Saudi Arabia's Vision 2030 objectives for a diversified and sustainable tourism sector.