Predictive classification modelling of sustainable tourism practices using online travel agent platform: a Malaysian case study

Tourism is a key pillar of Malaysia’s economy, recognized in the Economic Transformation Programme (ETP) as one of the 12 National Key Economic Areas (NKEA) that contribute to high-income status. However, the sector faces challenges, particularly in accommodation, due to environmental concerns and difficulties in measuring sustainable tourism. This study applies supervised machine learning classification techniques to develop a predictive model that classifies future sustainable tourism practices within Malaysia’s accommodation sector. Using data from an Online Travel Agent (OTA) platform, the model addresses data gaps and provides valuable insights for informed decision-making, aligned with the environmental conservation. The research is divided into four phases: literature review, data collection via web scraping, data modelling, and model evaluation. As the outcome, the multi-layer perceptron (MLP) model with 22 sets of features outperformed others, achieving an F1-score of 58%, recall of 83%, and Precision Recall Area Under the Curve (PRAUC) of 65%. These findings enhance both the understanding and forecasting of sustainability trends in tourist accommodation using online data.