HardMatch: Exact lexical string comparison

This metric aims to evaluate whether a generated label exactly matches any existing label in the gold-standard ontology. If the generated label matches a gold standard label (character-for-character), the result is 1; otherwise it is strictly 0.

Example:

Generate_Label_1: Wine

Generate_Label_2: wine

Generate_Label_3: Liquor

Gold_Standard_Label: Wine

Result:

  1. HardMatch(Generate_Label_1, Gold_Standard_Label) = 1
  2. HardMatch(Generate_Label_2, Gold_Standard_Label) = 0
  3. HardMatch(Generate_Label_3, Gold_Standard_Label) = 0

What is being measured?

This metric strictly compares the generated label to the gold-standard labels at the string level. It is the strictest matching approach with zero tolerance for variation and no semantic component.

Metric information

RDF Resource Description Framework Flyer Icon
Title
Exact lexical string comparison
Identifier
Keywords
Ontology Concept Matching, Metric, Classes, Properties, Lexical
Contact Point
Publisher
Version
0.1.0
License
Category
Ontology Concept Matching
Dimension
Task
Applicable for
Benchmark
CQ2OE Benchmark: Competency Questions for Ontology Engineering. A benchmark for evaluating LLM-assisted ontology generation from competency questions.
Status
Active
Active Proposed Inactive Deprecated
Ontology Engineering Group https://oeg.fi.upm.es