EEAT-Scorer: E-E-A-T Signal Scoring from Content and Links

Type: Academic | Domain: SEO, Trust & Quality
Hugging Face: syeedalireza/eeat-scorer

Score Experience, Expertise, Authoritativeness, and Trustworthiness signals from content and link data.

Author

Alireza Aminzadeh

• Hugging Face: syeedalireza
• LinkedIn: alirezaaminzadeh
• Email: alireza.aminzadeh@hotmail.com

Problem

E-E-A-T influences rankings and quality assessment. Automating signal extraction helps prioritize content and link improvements.

Approach

• Input: Page content (text), optional link metrics (referring domains, authority).
• Output: Scores or labels for experience, expertise, authoritativeness, trust (e.g. 0–1 or ordinal).
• Models: Sentence/document embeddings + regression or classification; optional transformer fine-tuned on labeled E-E-A-T data.

Tech Stack

Category	Tools
NLP	sentence-transformers, Hugging Face Transformers
ML	scikit-learn, XGBoost
Data	pandas, NumPy

Setup

pip install -r requirements.txt

Usage

python train.py
python inference.py --text "Article content here..." --referring_domains 50

Project structure

04_eeat-scorer/
├── config.py
├── train.py           # Content embeddings + XGBoost per E-E-A-T dimension
├── inference.py       # Score content (and optional link features)
├── requirements.txt
├── .env.example
├── data/
│   └── eeat_labels.csv   # Sample: content + E-E-A-T scores + optional referring_domains, dr
└── models/

Data

• Sample data (included): data/eeat_labels.csv — columns: content, experience, expertise, authoritativeness, trust (0–1); optional: referring_domains, dr.
• Set DATA_PATH in .env if using another file.

License

MIT.