A Deep Learning Approach to Comparative Sentiment Analysis for Ride-Hailing Apps

The swift growth of ride-hailing services in urban transportation has transformed the mobility landscape, necessitating that providers grasp customer sentiment to enhance their offerings. This research conducts a comparative sentiment analysis of user feedback for Ola, Uber, Rapido, and Namma Yatri, with the objective of deriving significant insights regarding customer satisfaction and service quality. A comprehensive dataset comprising 40,000 reviews from the Google Play Store for each application was gathered and classified into positive, neutral, and negative sentiments. To analyze sentiment trends, both advanced deep learning models (including LSTM, GRU, and Hybrid LSTM-CNN) and traditional machine learning models (such as Random Forest and Decision Tree) were utilized. The results show that even though the Random Forest model had the highest accuracy among the traditional methods, the Hybrid LSTM-CNN model was the best among all, reflecting the power of deep learning architectures in identifying complex sentiment patterns. The findings obtained from this study gave significant recommendations for ride-hailing companies to boost their customers’ satisfaction levels and streamline business processes. Keywords — Sentiment Analysis, Ola, Uber, Rapido, Namma Yatri, Machine learning, Deep learning, LSTM, GRU, Hybrid LSTM-CNN