Sentiment Analysis of Indian Tourist Place Reviews: A Machine Learning-Based Exploration

the social media trend, where many followers are interested in visiting the areas based on the reviews offered by the influencers, has caused the tourism industry to start growing incredibly high in recent years. This study conducts sentiment analysis on user-generated reviews from online tourism review websites, employing advanced feature extraction techniques. The dataset which has been taken from different web sources contains textual reviews and numerical ratings, providing the framework for analysis. Data preprocessing includes text tokenization, stop word removal, lemmatization, and data cleaning, resulting in a refined dataset with a processed review column. For feature extraction, we used TF-IDF vectorization (limited to 5000 features) and GloVe-based word embedding for enhancing word importance and semantic relationships. We employ different machine learning models such as Naive Bayes, SVM, Logistic regression, and Random Forest Classifier (with 100 decision trees) for sentiment analysis, which is designed to handle complex data relationships. Random Forest has given the best accuracy of 89.54 percent compared to the other models, while future work will encompass precision, recall, and F1-score. Our novel approach integrates TF-IDF, word embedding, and Random Forest for improved sentiment analysis. This methodology results of subsequent discussion by a robust foundation for further exploration in this field.