The Effect of Search Intent Data on Predicting Post-Crisis Daily Tourist Arrivals in Sri Lanka

Abstract

Accurately predicting daily international tourist arrivals is challenging for tourism-dependent economies like Sri Lanka, especially during its post-crisis recovery (2023–2025). Existing methods often ignore non-Google search engines, missing critical signals from key source markets like Russia, where Yandex dominates. Rather than proposing novel algorithmic architectures, this paper focuses on applied system integration and introduces an innovative data-model pipeline that integrates official daily arrivals with Yandex data, localized Google Trends, exchange rates, and weather records. We evaluate seven standard architectures, Pure SARIMA, SARIMAX, XGBoost, Random Forest, LSTM, Support Vector Regression (SVR), and a Sequential Hybrid, over a 550-day rolling test period. To prevent data leakage, feature selection utilizes predictive precedence and cross-correlation analysis strictly on training data. The results demonstrate that SVR with an RBF kernel performs best under our aligned feature framework, achieving a rolling MAPE of 8.94% and an RMSE of 736 arrivals/day. Crucially, Yandex queries for tours and flights demonstrate strong incremental predictive validity across all machine learning models, substantially outperforming Russian Google Trends data, which proved uninformative. These findings confirm that aligning predictive features both with a tourist market’s preferred search engine and the mathematical geometry of the underlying model significantly enhances forecasting accuracy in modern tourism intelligence systems.