Predicting Tourism Firms' Performance Indicators Using the Decision Tree Algorithms

Helhel Y., Yılmaz Y., Akgün A.

Advances in Hospitality and Tourism Research (AHTR), sa.Article in Press, ss.1-24, 2025 (ESCI, Scopus)

Özet

This study explores how well the most popular decision tree algorithms can predict the financial and stock return performances of tourism firms listed on the stock exchange in an emerging economy. The models incorporate tourism growth metrics, macroeconomic variables, and firm-level financial indicators as predictors. To this end, four widely used decision tree algorithms are applied to quarterly data spanning from the first quarter of 2012 to the fourth quarter of 2024. The results indicate that the Random Forest algorithm achieves the strongest predictive performance for ROA, ROE, and stock return performance, recording the highest scores across all key metrics. The firms’ liquidity level (quick ratio), asset size, and real exchange rate as a macroeconomic variable are the most significant determinants of ROA and ROE. Regarding stock return performance as a market-based measurement, liquidity level and the real exchange rate remain critical, while tourism expenditures and the macroeconomic indicator of M2 money supply also prove to be significant determinants.

This study explores how well the most popular decision tree algorithms can predict the financial and stock return performances of tourism firms listed on the stock exchange in an emerging economy. The models incorporate tourism growth metrics, macroeconomic variables, and firm-level financial indicators as predictors. To this end, four widely used decision tree algorithms are applied to quarterly data spanning from the first quarter of 2012 to the fourth quarter of 2024. The results indicate that the Random Forest algorithm achieves the strongest predictive performance for ROA, ROE, and stock return performance, recording the highest scores across all key metrics. The firms’ liquidity level (quick ratio), asset size, and real exchange rate as a macroeconomic variable are the most significant determinants of ROA and ROE. Regarding stock return performance as a market-based measurement, liquidity level and the real exchange rate remain critical, while tourism expenditures and the macroeconomic indicator of M2 money supply also prove to be significant determinants.