Database of Paradigmatic Semantic Relation Pairs

The database is a collection of semantically related word pairs in German which was compiled via human judgement experiments hosted on Amazon Mechanical Turk. We address the three paradigmatic relations antonymy, hypernymy and synonymy. The database consists of three parts:

• A representative selection of target lexical units drawn from GermaNet, using a principled sampling technique and taking into account the three major word classes adjectives, nouns, and verbs, which are balanced according to semantic category, polysemy, and type frequency.
• A set of 8,910 human-generated semantically related word pairs, based on the target lexical units.
• A subset of 1,684 semantically related word pairs, rated for the strengths of relations.

Our dataset fulfils the following criteria:

• It focuses on multiple paradigmatic relations.
• It systematically works across word classes.
• It explicitly balances the targets according to semantic category, polysemy and type frequency.
• It explicitly provides positive and negative rating evidence.

Other datasets with paradigmatic relation pairs can be found here (published at ACL 2016) and here (published at EACL 2017).

Silke Scheible, Sabine Schulte im Walde (2014)
A Database of Paradigmatic Semantic Relation Pairs for German Nouns, Verbs, and Adjectives
In: Proceedings of the COLING Workshop on Lexical and Grammatical Resources for Language Processing (LG-LP). Dublin, Ireland.

Sabine Schulte im Walde
Distinguishing between Paradigmatic Semantic Relations across Word Classes: Human Ratings and Distributional Similarity
Journal of Language Modelling 8(1):53-101.

Please contact the SemRel group to obtain the data.

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA).