Postag

Overview

Postag is a simple part of speech tagger which is capable of two modes, a naive baseline mode and a second order Hidden Markov Model mode.

Usage

postag.py [mode] [train_data] [test_data]

• mode is either baseline which tags words with their most common part of speech from the training data or hmm which uses a second order Hidden Markov Model
• train_data and test_data must be of the same format as the included train.txt and test.txt

Files

All data from the SUSANNE corpus, compiled by Gregory Sampson et al.

• train.txt: Sample training data
• test.txt: Sample test data
• tagset.txt: Explanation of the part-of-speech tags used