| license: mit | |
| language: | |
| - en | |
| tags: | |
| - sentiment-analysis | |
| - imdb | |
| - scikit-learn | |
| - mlp | |
| - text-classification | |
| pipeline_tag: text-classification | |
| # IMDB Sentiment MLP | |
| This model is a course project for IMDB movie-review sentiment classification. | |
| It uses a TF-IDF text representation followed by a small `scikit-learn` MLP neural network. | |
| ## Metrics | |
| - Accuracy: 82.00% | |
| - Train samples: 400 | |
| - Test samples: 100 | |
| - Dataset: `imdb_top_500.csv` | |
| - Labels: `0 = negative`, `1 = positive` | |
| ## Files | |
| - `model.joblib`: full scikit-learn pipeline | |
| - `vectorizer.joblib`: standalone TF-IDF vectorizer | |
| - `metrics.json`: training and evaluation metrics | |
| ## Example | |
| ```python | |
| import joblib | |
| model = joblib.load("model.joblib") | |
| prediction = model.predict(["This movie is great and deeply moving."])[0] | |
| print("positive" if prediction == 1 else "negative") | |
| ``` | |