metadata
datasets:
- dpmendez/environmental-misinformation
language:
- en
This model is a logistic regression classifier trained on TF-IDF representations of climate-related claims. It serves as a transparent, interpretable baseline for climate misinformation detection.
The model is trained on a dataset combining:
- Climate Fever
- Science Feedback fact-checked claims and predicts the veracity of individual climate-related claims.
Model details
- Model type: Logistic Regression
- Input: Single climate-related claim (text)
- Output: Claim label
- Feature extraction: TF-IDF (unigrams + bigrams)
- Framework: scikit-learn
Labels
| Label | Description |
|---|---|
LIKELY_TRUE |
Claim is consistent with scientific consensus |
LIKELY_FALSE |
Claim contradicts scientific consensus |