A Unified POS Tagging Architecture and its application to Greek (original) (raw)

This paper presents a unified architecture for Part of Speech (POS) tagging that addresses the intricacies of tagging the Greek language, characterized by a complex morphosyntactic structure and a rich tagset. By integrating several components including a tokenizer, graphical annotation tool, and a feature-based multi-tiered tagger, this framework overcomes common challenges encountered in POS tagging, particularly for highly inflective languages. The result showcases improved accuracy and functionality in processing and annotating a diverse set of texts.