Learning to Case-Tag Modern Greek Text

Morphological case tagging is essential for the identification of the syntactic and semantic roles of sentence constituents in most inflectional languages. Although it is usually viewed as a side-task of general tagging applications, it is addressed in the present work as an individual, stand-alone application. Supervised learning is applied to Modern Greek textual data in order to case-tag declinable words using merely elementary lexical information and local context. Several experiments with various context window sizes, as well as base- and meta-learning schemata, were run with promising results.