Evaluating the Impact of Data Partitioning Strategies on Model Generalizability: A Cross-Linguistic Study of Morphological Segmentation

Different data partitioning strategies, such as random and adversarial splits, can significantly impact the generalizability of computational models for morphological segmentation across diverse languages.