Skip to main content
Version: 1.0.0

Glossary

This glossary collects the main concepts and acronyms used in MantisTable UI and in Semantic Table Interpretation (STI).

CEA (Cell Entity Annotation)​

The task of linking individual table cells to the most appropriate entity in a knowledge graph.
Example: the cell "Mont Blanc" in a mountain column is linked to the entity Mount_Blanc in the knowledge graph.

CPA (Column Property Annotation)​

The task of identifying the semantic relations between two columns in a table.
Example: linking the column "Name" and "Height" through the property dbo:elevation, or "Name" and "Coordinates" through georss:point.

CTA (Column Type Annotation)​

The task of assigning the most suitable class or type from a knowledge graph to a column.
Example: the column "Name" is annotated with the type Mountain, or "Range" with Mountain_Range.

Datatype annotation​

The identification of a column as containing literal values such as numbers, strings or dates, without linking to an external entity.
Example: a column "Height" annotated as xsd:integer, or "Coordinates" as xsd:string.

Entity linking​

The process of mapping a textual mention (e.g. a string in a table cell) to an unambiguous entity in a knowledge graph.
This is the underlying technique for CEA.

Knowledge Graph (KG)​

A structured representation of entities, classes and relations, usually expressed as RDF triples.
It provides the background knowledge for STI tasks.

Literal (LIT) column​

A column whose values are literals, i.e. not linked to external entities.
Example:

  • "Height" → values such as 4808, 3275 are treated as integers.
  • "Coordinates" → values such as 45° 49′ 57″ N, 06° 51′ 54″ E are treated as strings.

Named Entity (NE) column​

A column whose values are named entities, i.e. references to real-world objects that can be linked to a knowledge graph.
Example:

  • "Name" → values such as Mont Blanc, Cervino are linked to entities Mount_Blanc, Mount_Cervino.
  • "Range" → values such as M. Blanc massif, Pennine Alps are linked to entities Mount_Blanc_Massif, Pennine_Alps.

STI (Semantic Table Interpretation)​

The process of enriching tabular data with semantic annotations derived from knowledge graphs.
It typically includes CEA, CTA, and CPA tasks.

Why the NE/LIT distinction matters​

Distinguishing between Named Entity (NE) and Literal (LIT) columns is crucial for semantic table interpretation.
It determines whether cells should be linked to knowledge graph entities (NE) or treated as raw values (LIT) with an appropriate datatype.
This distinction affects the accuracy of CEA, the choice of relations in CPA, and the utility of the resulting annotations in downstream applications.