Introduction
This project constitutes 30% of your final grade for HG2051. Please work on the final program in groups of 2-3 and report together.
The goal of this assignment is to demonstrate your programming and problem-solving abilities through teamwork. If your team has an idea for another project that you would like to do instead, talk to me for approval.
This project involves training a part of speech tagger based on existing tagged data and using it to tag unseen data. As with the individual project, your team will be required to submit data, output, and annotated code along with a short writeup that describes your goals, process, and results. Your code will be assessed based on its functionality and simplicity.
Project 2: Part of speech tagging
Part of speech tagging is a way to automatically identify the word class of a particular lexical item in a string of text. A decent part of speech tagger can help to facilitate other downstream tasks such as machine translation. The initial project code in the repository will help you to get started, along the following lines:
Two text files are provided for this project, corresponding to the two files you used for Project 1 (if you chose to work on the language I provided files for). The first text will be used to train your POS-tagger, and the second will be used to validate it.
Parse the first text into word-POS tuples and use them to train a POS-tagger. This can be done using the NLTK library (see NLTK Ch 5), but you may want to use a different library. Test your POS-tagger on the second text to see how well it performs - this second text has been edited to include only POS-tagged sentences. The metric you will use to determine performance is
accuracy.The provided script gives a baseline accuracy using modules in the NLTK library. Your accuracy should improve once you add more data to train your POS-tagging model. The goal of your project is to improve on the existing baseline, by additionally including rules or features for your POS-tagger. To explore potential rules/features, you may want to do some research on the language itself, or find some other methods for developing POS-taggers.
In order to improve your POS-tagger further, consider whether to incorporate additional data from more sources, such as those linked in Project 1. You may also want to train multiple POS-taggers and use them in combination to see if you can improve your score.
This code is simply a starting point for exploration. What other features can you think of and implement to improve accuracy? Are there other lexical resources/corpora or algorithms within NLTK or from other sources that can improve your results? What is the best accuracy you can get? How relevant are these POS tags to other languages, i.e. would it be good to try and convert them to something like the Universal Tag Set?
With your group, write a roughly 5-page paper (no longer than 10 pages) describing your goals, data, process, and results, and discussing some of the concerns identified in point 5. This paper should include relevant linguistic information and citations for your sources. Submit the paper as a PDF along with your annotated/commented Python code in a Github repository. The PDF should also be submitted via TurnItIn.