Machine Learning Forecasting Techniques for Analyzing Tourist Arrivals in Bali

Abstract

This study investigates the application of machine learning (ML) techniques for forecasting tourist arrivals in Bali, leveraging a dataset spanning from 1982 to 2024. The Random Forest model, along with Linear Regression and Decision Tree, was evaluated for its ability to handle the complexities of tourism data, characterized by seasonality and nonlinear patterns. Among the models tested, Random Forest achieved the best performance, with the lowest Mean Squared Error (MSE) and Mean Absolute Error (MAE), demonstrating its robustness in capturing both short-term fluctuations and long-term trends. The findings highlight the potential of ML techniques to improve forecasting accuracy compared to traditional methods, especially in managing seasonal variations and external disruptions like the COVID-19 pandemic. However, limitations in predicting unprecedented events underscore the need for integrating external variables, such as economic indicators and travel restrictions. Future research should focus on hybrid models, scenario-based forecasting, and real-time data integration to enhance adaptability and predictive accuracy. These advancements aim to support policymakers and stakeholders in optimizing resource allocation, designing marketing strategies, and fostering sustainable tourism development in Bali.