------------------------------------------------------------------------------------------------------------------------
6320.501: Natural Language Processing
------------------------------------------------------------------------------------------------------------------------
Author Name: Dhwani Raval
Project Name: Brill's Tagger
Version: 0.1
Programming Language: Python 3.6
------------------------------------------------------------------------------------------------------------------------
Problem Description
------------------------------------------------------------------------------------------------------------------------
Transformation-based POS Tagging: Implement Brill’s transformation-based POS tagging algorithm using ONLY the previous
word’s tag to extract the best five (5) transformation rules to:
1. Transform “NN” to “VB”
2. Transform “VB” to “NN”
Using the learnt rules, manually fill out the missing POS tags (for the word “control”) in the following sentence:
The_DT president_NN wants_VBZ to_TO control_??? the_DT board_NN 's_POS control_???
------------------------------------------------------------------------------------------------------------------------
About Files
------------------------------------------------------------------------------------------------------------------------
The project contains the following files:
1. sourcecode/Tagger.py: The python file for the given problem description
2. resources/POSTaggedTrainingSet.txt: A training set that has been tagged with POS tags from the Penn Treebank POS
tagset
3. output/tuple: A text file created during program execution
4. output/unigram: Text files related to unigrams created during program execution
5. output/tags: Text files related to correct and current tags for words created during program execution
6. output/logs: Log files created during each iteration for top 10 rules
7. output/top10.txt: Top 10 transformation rules
8. readme: A text file containing information about the project
------------------------------------------------------------------------------------------------------------------------
Brill's Tagging Description
------------------------------------------------------------------------------------------------------------------------
Brill's Tagging is used for Part-of-Speech (POS) tagging. It is inductive in nature and is based on
Transformation Based Learning (TBL). The basic idea is to iteratively assign the best tag to a word
using the learned transformations based upon a set of predefined rules. The goal of this approach is to
minimize the error rate in every step.
------------------------------------------------------------------------------------------------------------------------
Running Instructions
------------------------------------------------------------------------------------------------------------------------
1. Download the project and unzip it in the desired location
2. In IDE, import the project and run Tagger.py
3. In cmd, navigate to the location where Brill_Tagging is unzipped and use the following instruction
python sourcecode\Tagger.py
dhwaniraval/Brill_Tagger
Transformation-based POS Tagging: Implemented Brill’s transformation-based POS tagging algorithm using ONLY the previous word’s tag to extract the best five (5) transformation rules to: 1. Transform “NN” to “VB” 2. Transform “VB” to “NN”
Python