Published March 6, 2019 | Version 0.1.1
Dataset Open

A labeled Ecore metamodel dataset for domain clustering

  • 1. Eindhoven University of Technology


Manually labeled 555 metamodels mined from GitHub in April 2017. 

Domains: (1) bibliography, (2) conference management, (3) bug/issue tracker, (4) build systems, (5) document/office products, (6) requirement/use case, (7) database/sql, (8) state machines, (9) petri nets

Procedure for constructing the dataset: fully manual, by searching for certain keywords and regexes (e.g. "state" and "transition" for state machines) in the metamodels and inspecting the results for inclusion. 

Format for the file names: ABSINDEX_CLUSTER_ITEMINDEX_name_hash.ecore


Files (1.0 MB)

Name Size Download all
1.0 MB Preview Download