A system that uses multiple approaches (regex patterns, embeddings, and machine learning) to classify log messages efficiently.
- Log message clustering using DBSCAN and sentence embeddings
- Rule-based classification using regex patterns
- ML-based classification using LogisticRegression on embeddings
- Special handling for low-data sources