Deep Learning and Review Sentiment Analysis for Hotel Recommendations System

Travelers often rely on online reviews to choose the best hotel when they are traveling outside of their home nation or region due to an increasing number of online travel platforms. However, travelers may find it overwhelmed by the overload, and various reviews pose a challenge in effectively sorting through the opinions of specific hotels. This study compares several deep learning techniques, including convolutional neural networks (CNN), long short-term memory (LSTM), and bidirectional encoder representations from transformers (BERT), to determine which is the most efficient and accurate for sentiment analysis of hotel reviews. Our findings demonstrate BERT's superiority, achieving an accuracy of 86%. This shows BERT's ability to capture the context and emotion of words within a sentence, leading to more accurate sentiment analysis.