Python Natural Language Processing

Developing POS taggers step-by-step

To build your own POS tagger, you need to perform the following steps:

1. You need a tagged corpus.
2. Select features.
3. Perform training using a decision tree classifier available in the Python library, scikit-learn.
4. Check your accuracy.
5. Try to predict the POS tags using your own trained model.

A great part of this section is that we will code our own POS tagger in Python, so you guys will get an idea of how each of the preceding stages are performed in reality. If you don't know what a decision tree algorithm is, do not worry - we will cover this topic in more detail in Chapter 8, Machine Learning for NLP Applications.

Here, we will see a practical example that will help you understand the process of developing ...