A variety of Natural Language Processing and Information Extraction tasks, such as question answering and named entity recognition, can benefit from precise knowledge about a words' syntactic category or Part of Speech (POS) (Church, 1988; Rabiner, 1989; Stolz, Tannenbaum, & Carstensen, 1965). POS t aggers are widely used to assign a single best POS to every word in text data, with stochastic approaches achieving accuracy rates of up to 96% to 97% (Jurafsky & Martin, 2000). When building a POS tagger, human beings need to make a set of choices about design decisions, some of which significantly impact the accuracy and other performance aspects of the resulting engine. However, documentations of POS taggers often leave these choices and decisions implicit. In this paper we provide an overview on some of these decisions and empirically determine their impact on POS tagging accuracy. The gained insights can be a valuable contribution for people who want to design, implement, modify, fine-tune, integrate, or responsibly use a POS tagger. We considered the results presented herein in building and integrating a POS tagger into AutoMap, a tool that facilitates relation extraction from texts, as a stand-alone feature as well as an auxiliary feature for other tasks.

35 pages

Return to: SCS Technical Report Collection
School of Computer Science homepage

This page maintained by reports@cs.cmu.edu