|
|
---
|
|
|
tags:
|
|
|
- nlp
|
|
|
- regression
|
|
|
- tfidf
|
|
|
- ridge
|
|
|
- summaries
|
|
|
- kaggle
|
|
|
---
|
|
|
|
|
|
# 🧠 CommonLit Summary Scoring Model
|
|
|
|
|
|
This model was trained using the **CommonLit Evaluate Student Summaries** dataset on Kaggle.
|
|
|
It predicts two scores for student-written summaries:
|
|
|
|
|
|
- `content` → Idea coverage quality
|
|
|
- `wording` → Clarity and phrasing quality
|
|
|
|
|
|
Built with:
|
|
|
- TF-IDF vectorizer
|
|
|
- Ridge Regression (scikit-learn)
|
|
|
- MultiOutputRegressor wrapper
|
|
|
|
|
|
Example usage:
|
|
|
```python
|
|
|
from joblib import load
|
|
|
|
|
|
model = load("ridge_model.pkl")
|
|
|
tfidf = load("tfidf_vectorizer.pkl")
|
|
|
summary = "This text discusses..."
|
|
|
X = tfidf.transform([summary])
|
|
|
pred = model.predict(X)
|
|
|
|
