metadata

license: mit
task_categories:
  - text-classification
language:
  - ka
size_categories:
  - 1M<n<10M

Georgian-Homonym-Disambiguation

This repository contains all the datasets for the Georgian homonym disambiguation task.

For more specific details you can read my article

Dataset

At this point I've considered only the homonym: "ბარი" and it's different grammatical forms obtaining 7522 sentences.

The "dataset.parquet" includes:

763 sentences using "ბარი" as a "shovel" labaled with 0
1846 sentences using "ბარი" as a "lowland" labeld with 1
3320 sentences using "ბარი" as a "cafe" labeled with 2
1593 sentences where the homonym is used in a different context, labeled with 3 (Although these sentences could be further classified by the definitions of the homonyms, for this project I've ignored other usages).

the column 'homonym_index' contains the index of the homonym in the sentence, that is, the index of the word which is the homonym.

The "full-homonym-sentences-ბარ.txt" includes the sentences which contain the homonym "ბარი" and it's various grammatical forms. These sentences were limited to a maximum length of 13 words, with the homonym positioned in the middle of each sentence. They are around 28000 and are not labelled.