End-to-End pipeline to identify matching records referring to the same publication.

This project uses two bibliographic datasets (DBLP1.csv, Scholar.csv) and a ground-truth mapping (DBLP-Scholar_perfectMapping.csv).

Run the complete pipeline:

python run.py

This command will run all the notebooks with papermill. The notebooks will save their results in .json files within their folders.

• T1: Data Preperation ✓
• T2: Blocking ✓
• T3: Similarity Scoring ✓

• T1: Matching & Evaluation ✓
• T2: Training Dataset ✓
• T3: Feature Engineering ✓

• T1: Training Data ✓
• T2: Model Training ✓
• T3: Evaluation ✓