Welcome to NLTK-Trainer’s documentation!

NLTK-Trainer is a set of Python command line scripts for natural language processing. With these scripts, you can do the following things without writing a single line of code:

1. train NLTK based models
2. evaluate pickled models against a corpus
3. analyze a corpus

These scripts are Python 2 & 3 compatible and work with NLTK 2.0.4 and higher.

Download

The scripts can be downloaded from nltk-trainer on github.

Documentation

• Training Classifiers
• Using a Trained Classifier
• Training Part of Speech Taggers
• Using a Trained Tagger
• Training IOB Chunkers
• Using a Trained Chunker
• Analyzing a Tagged Corpus
• Analyzing Tagger Coverage

Books

Python 3 Text Processing with NLTK 3 Cookbook contains many examples for training NLTK models with & without NLTK-Trainer.

• Chapter 4 covers part-of-speech tagging and train_tagger.py.
• Chapter 5 shows how to train phrase chunkers and use train_chunker.py.
• Chapter 7 demonstrates classifier training and train_classifier.py.

Articles

• Training Binary Classifiers with NLTK Trainer
• Training Part of Speech Taggers with NLTK Trainer
• Analyzing Tagger Corpora and NLTK Part of Speech Taggers
• NLTK Default Tagger Coverage of treebank corpus
• NLTK Default Tagger Coverage of conll2000 corpus

Demos and APIs

Nearly all the models that power the text-processing.com NLTK demos and NLP APIs have been trained using NLTK-Trainer.

Indices and tables

• Index
• Module Index
• Search Page