Expand description
Error analysis and diagnostics for NER evaluation.
Provides tools to understand why a model makes mistakes:
- Confusion matrix (which types get confused)
- Error categorization (boundary, type, spurious, missed)
- Statistical significance testing between systems
Structs§
- Confusion
Matrix - Confusion matrix for entity type predictions.
- Error
Analysis - Error analysis results.
- NERError
- A single NER error instance.
- NERSignificance
Test - Result of paired significance test for NER systems.
Enums§
- Error
Type - Categories of NER errors.
Functions§
- build_
confusion_ matrix - Build confusion matrix from predictions and ground truth.
- compare_
ner_ systems - Compare two NER systems with significance testing.