Extraction Methodology

Temporal expressions are extracted using a hybrid approach combining spaCy's Named Entity Recognition (DATE entities) with rule-based pattern matching for years, decades, centuries, historical periods, and relative temporal markers. Where possible, expressions are normalized to ISO 8601 format or standardized date ranges.

Future enhancements: Subsequent versions will incorporate LLM-based validation to improve extraction precision, contextual disambiguation for ambiguous expressions, and temporal relation extraction to identify event ordering and duration relationships within the document.