Zero-shot cross-lingual transfer language selection using linguistic similarity

We study the selection of transfer languages for different Natural Language Processing tasks, specifically sentiment analysis, named entity recognition and dependency parsing. In order to select an optimal language, we propose utilize linguistic similarity metrics measure distance between make choice language based on this information instead relying intuition. demonstrate that correlates with cross-lingual performance all proposed tasks. also show there is a statistically significant difference in choosing as source English. This allows us more suitable which can be used better leverage knowledge from high-resource improve applications lacking data. For study, datasets eight three families.