Sentiment Analysis of Hotel Customers' Feedback in Bangla Using Machine Learning

In the rapidly expanding internet era, customers or tourists can easily find information about hotels. The growth of the hospitality industry in Bangladesh has led to a rise in the number of online hotel reviews. In this respect, it becomes a concern to understand the customer experience and review the quality level of hotels relying on their services. Sentiment analysis is one of the intimate extents of text mining, where digital text-based consumer responses or emotions are reviewed using machine learning methods. The purpose of this study is to construct a benchmark dataset of hotel customers' feedback in Bangla and preprocess the data for sentiment analysis using natural language processing techniques. In accordance with this, it has proposed a machine learning method to accomplish sentiment analysis on Bangla text-based hotel reviews. It has been employed a set of machine learning algorithms like Decision Tree (DT), Random Forest (RF), Support Vector Machines (SVM), and Naive Bayes (NB) to explore the accuracy of Bangla text data regarding customer feedback. These algorithms classify the reviews as positive, negative, or neutral. This study performs these algorithms and finds the consequences of machine learning models in terms of accuracy. However, the proposed analysis from these algorithms shows that RF provides the highest accuracy of 83% for sentiment analysis tasks. These findings suggest that combining machine learning techniques can be an effective way to analyze Bangla text-based hotel customers' feedback and provide valuable insights for the hospitality industry in Bangladesh.