This project focuses on analyzing the sentiment of IMDB movie reviews (positive or negative) using a Simple Recurrent Neural Network (RNN). The model processes the sequential nature of text data to predict the sentiment of a given review.
- Sentiment analysis using Simple RNN.
- Preprocessing of IMDB movie review data.
- Model evaluation with accuracy and loss metrics.
- Visualizations of training performance (loss and accuracy).
- Easy-to-follow code for learning and replication.
- Accuracy: ~85% on the test set.
- Loss: Tracked over 10 epochs.
Metric | Value |
---|---|
Train Accuracy | 86% |
Test Accuracy | 84% |
Epochs | 10 |
- Use more advanced architectures like LSTM, GRU, or Transformers.
- Explore data augmentation for better performance.
- Add a web-based interface for live predictions.
- Experiment with hyperparameter tuning for improved accuracy.