Comparative Analysis of Deep Learning Models for Sentiment Classification of Indian Automobile YouTube Comments

The rapid expansion of the Indian automotive industry has resulted in a rise in user-generated material on websites such as YouTube. Examining these viewpoints provides manufacturers, dealerships, and legislators with insightful information. A sentiment analysis framework for dividing YouTube comments about Indian cars into three categories - positive, neutral, and negative - is proposed in this paper. Microsoft Power Automate was used to create an automated pipeline that gathered and cleaned 5,203 comments in total, guaranteeing effective and organized data extraction. Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and Bidirectional Encoder Representations from Transformers (BERT) are the four deep learning models that were used and assessed.BERT outperformed the other models in classification performance, with the maximum accuracy of 88.7% and an F1-score of 0.86. The paper talks about the trade-offs between accuracy, training time, and model complexity. For the implementation of sentiment analysis in actual automotive applications, this framework provides useful insights. Future research will examine real-time monitoring systems and multilingual sentiment classification.