Sentiment Analysis on YouTube Comments using Long Short-Term Memory (LSTM) Networks

This research study analyzes YouTube comments' sentiment using Long Short-Term Memory (LSTM) networks, it is a type of Recurrent Neural Network (RNN) effective for sequential data. Our aim is to classify comments into positive and negative categories. Considering the rise in content creators on different social media platforms, understanding comment sentiment is vital for them. We gathered a dataset of YouTube comments, preprocessed it to handle slang, abbreviations, and emoticons. Using word embeddings and LSTM layers, our model keeps contextual information and outperforms traditional machine learning algorithms and other deep learning architectures in accuracy, precision, and recall. Practical applications include enhanced content moderation, optimized recommendation systems, and real-time feedback for creators. Our findings highlight the importance of continuous model refinement and multimodal data integration for future sentiment analysis research.