STROMA (Semantic Refinement of Ontology Mappings) is a mapping enrichment tool that calculates the relation types of the correspondences within a given schema or ontology mapping. The knowledge about the concrete relation type, e.g., equal, is-a, has-a or part-of is very substantial in data integration and related fields, such as ontology merging, data transformation or ontology evolution. To determine the relation types, STROMA exploits different linguistic strategies, combined with background knowledge, structual strategies and some linguistic heuristics.

Image

The program requires a textual mapping as input and produces an enhanced mapping with the semantic types denoted at each correspondence. The required input format is simplicity itself; it is a CSV-like text file with each line representing one correspondence. For each correspondences, the source concept path and target concept path must be available, and optionally a confidence value (score). Additionally, STROMA is also able to load and process mappings created with COMA 3.0.

In order to obtain optimized results, STROMA uses the background knowledge repository SemRep to determine the relation types, including versatile resources like WordNet, UMLS, ConceptNet and automatically extracted concept relations from Wikipedia.

Download

STROMA is available under an AGPL license.

• Click here to download STROMA.
• Click here to download the STROMA User’s Guide.

A benchmark including evaluation data and documentation is available.

• Click here to download the benchmark.