Analyzing Tagging Accuracy of Part-of-Speech Taggers

Automated part-of-speech (POS) tagging has been a very active research area for many years and is the foundation of natural language processing systems. Natural Language Toolkit (NLTK) library in the Python environment provides the necessary tools for tagging, but doesn’t actually tell us what methods work the best. Therefore, this work analyzes the performance of part-of-speech taggers, namely the NLTK Default tagger, Regex tagger and N-gram taggers (Unigram, Bigram and Trigram) on a particular corpus. The corpora we have used for the analysis are; Brown, Penn Treebank and CoNLL2000. We have applied all taggers to these three corpora, resultantly we have shown that whereas Unigram tagger does the best tagging in all corpora, the combination of taggers does better if it is correctly ordered.

Authors and Affiliations

1. University of Computer Studies, Mandalay, Myanmar

   Nyein Pyae Pyae Khin & Than Nwe Aung

Corresponding authors

Correspondence to Nyein Pyae Pyae Khin or Than Nwe Aung .

Editors and Affiliations

1. Faculty of Engineering, University of Miyazaki, Miyazaki, Japan

   Thi Thi Zin

2. School of Computer Science and Tech..., Harbin Institute of Technology, Shenzhen Graduate School , Shenzhen, China

   Jerry Chun-Wei Lin

3. College of Information Science and Engg, Fujian University of Technology, Fuzhou, China

   Jeng-Shyang Pan

4. Faculty of Engineering, University of Miyazaki, Miyazaki, Japan

   Pyke Tin

5. Faculty of Engineering, University of Miyazaki, Miyazaki, Japan

   Mitsuhiro Yokota

Reprints and permissions

© 2016 Springer International Publishing Switzerland

Policies and ethics