Sentiment Analysis of Online Hotel Reviews Employing Bidirectional GRU with Attention Mechanism

Online hotel reviews are a more reliable resource for potential hotel guests. Sentiment analysis is a branch of text mining, Natural Processing Language that seeks to identify personality traits, emotions, and other factors. Deep Learning algorithms such as LSTM and GRU have successfully generated context information in sequence learning. However, deep learning cannot focus on the words that contribute the most and cannot capture important content information. This research aims to overcome the inability of LSTM and GRU to capture information. The results are satisfactory, with 93.12% accuracy, 95% ROCAUC, and 95.28% precision recall. This research paper helps managers identify areas to improve their products and services, target marketing campaigns, and identify customer churn.