Rankings of the output of distributional semantic models of Ancient Greek

This dataset contains an evaluation of the output of distributional semantic models, as described in: Keersmaekers, Alek & Dirk Speelman. 2023. Applying Distributional Semantic Models to a Historical Corpus of a Highly Inflected Language: the Case of Ancient Greek. Glottometrics 55. 17–43. The 10 nearest neighbors of 100 Ancient Greek lemmas, as returned by 5 distributional semantic models, were assigned a semantic category depending on the relation of the target word and its neighbor. The semantic categories roughly correspond to common categories in the scientific literature on the evaluation of word embeddings: 'synonym': synonymous 'related': similar (we were not aware of the similar-related distinction that is common in the literature when this paper was written) 'distantly related': distantly similar 'same domain': related 'unrelated': unrelated See the paper cited above for more details. The data used in the paper can be found in column R2: these data were annotated by Toon Van Hal. Additionally, column R1 was annotated by Alek Keersmaekers.