Sentiment Analysis of Turkish Hotel Reviews: A Tripadvisor Study Case

Sentiment analysis helps users understand the overall sentiment towards services quickly and effectively. Especially after the pandemic, consumers increasingly rely on online reviews to evaluate products and services. However, the abundance of available information makes it difficult for users to examine each review. In this study, sentiment classification of Turkish hotel reviews on TripAdvisor was conducted using machine learning, deep learning, and transformer-based architectures, and Turkish reviews were analyzed. In addition, the effectiveness of hybrid CNN-LSTM and CNN-GRU deep learning models was compared with transformer methods. The results obtained showed that hybrid deep learning architectures and transformerbased architectures achieved over 90% success. In addition to these state-of-the-art architectures, LR and SVM machine learning algorithms in particular also yielded very high results in sentiment analysis. The results obtained can offer innovative service development suggestions based on customer feedback for hotel businesses.