Evaluating Part-of-Speech Tagging and Parsing Patrick Paroubek

The aim of this chapter is to introduce the reader to the evaluation of part-ofspeech (POS) taggers and parsers. After a presentation of both POS tagging and parsing, describing the tasks and the existing formalisms, we introduce general considerations about evaluation of Natural Language Processing (NLP). Then we raise a point about the issue of input data segmentation into linguistic units, a crucial step in any evaluation related to language processing. We conclude by a review of the current evaluation methodologies and average levels of performance generally achieved for POS tagging and parsing.