Natural Language Processing: Python and NLTK

Training a tagger with NLTK-Trainer

As you can tell from all the previous recipes in this chapter, there are many different ways to train taggers, and it's impossible to know which methods and parameters will work best without doing training experiments. But training experiments can be tedious, since they often involve many small code changes (and lots of cut and paste) before you converge on an optimal tagger. In an effort to simplify the process, and make my own work easier, I created a project called NLTK-Trainer.

NLTK-Trainer is a collection of scripts that give you the ability to run training experiments without writing a single line of code. The project is available on GitHub at https://github.com/japerk/nltk-trainer and has documentation ...