CroWSI - Graph-Based Induction of Word Senses in Croatian

1 Description

CroWSI is a set of co-occurrence graph-based WSI experiments for Croatian.

2 Dataset

The dataset is available from here: TakeLab-CroWSI.zip.

The archive contains three files:

• first containing an gold standard evaluation dataset (evaluation_dataset.json),
• second containing a fine-grained inducted sense inventory for the 10,000 most frequent Croatian words (fine_grained_sense_inventory.json),
• and the third containing a coarse-grained inducted sense inventory for the 10,000 most frequent Croatian words (coarse_grained_sense_inventory.json)

3 License