Bi-GRU and Glove Based Aspect-Level Movie Recommendation

This study investigates the effectiveness of deep learning architectures for sentiment analysis in text-based data. The methodology begins with data collection, where movie reviews sourced from IMDb are utilized, reflecting a variety of emotional expressions. The dataset undergoes preprocessing, which includes tokenization, embedding using GloVe, and splitting into training and testing sets, with 80 allocated for training and 20 for testing to ensure reliable model evaluation. The models implemented in this study include GRU with GloVe embeddings, BiGRU with GloVe embeddings, and LSTM with GloVe embeddings. Each model is constructed to capture sequential dependencies in the text, allowing for a nuanced understanding of sentiment. The models are compiled with appropriate optimizers and loss functions, and their performance is evaluated using accuracy, precision, and recall metrics. During the testing phase, the trained models are assessed based on their ability to classify sentiments effectively, generating insights into their performance. The results are visualized to highlight misclassification and areas for improvement. This methodology streamlines the sentiment classification process, reducing reliance on manual analysis and enhancing overall efficiency. The findings demonstrate that deep learning models can significantly improve sentiment analysis accuracy, providing valuable insights for future research in emotion detection within Natural Language Processing (NLP) applications.