How shivambharti-29/sentiment-analysis-rnn-streamlit works
This is a Streamlit movie-review sentiment app. A user submits a review; the app cleans the text, vectorizes it with TF-IDF, classifies it, and displays a positive/negative result with a confidence score. The README documents the prediction stages; the excerpt confirms the app loads the serialized model and vectorizer and imports the cleaner. The prediction-call details are partly omitted, so those pipeline links are documented relationships rather than inspected call sites.
Main entry points: app.py Review App · preprocessing.py Text Cleaning · vectorizer.pkl TF-IDF Vectorizer · model.pkl Serialized Model
Components 7
User Interface
- Review App
app.py - Sentiment Result
app.py
Sentiment Inference
- Text Cleaning
utils/preprocessing.py - TF-IDF Vectorizer
model/vectorizer.pkl - RNN Classifier
app.py
Model Artifacts
- Serialized Model
model/model.pkl
Other components
- Review Author
Connections 6
- Review Author → Review App: submits review
README.md - Review App → Text Cleaning: cleans text
app.py - Review App → TF-IDF Vectorizer: vectorizes text
README.md - Review App → RNN Classifier: classifies text
README.md - Review App → Serialized Model: loads model
app.py - Review App → Sentiment Result: displays prediction
README.md