Converting tree labels
by Iti Mathur, Nisheeth Joshi, Deepti Chopra, Jacob Perkins, Nitin Hardeniya
Natural Language Processing: Python and NLTK
- Natural Language Processing: Python and NLTK
- Table of Contents
- Natural Language Processing: Python and NLTK
- Natural Language Processing: Python and NLTK
- Credits
- Preface
- 1. Module 1
- 1. Introduction to Natural Language Processing
- 2. Text Wrangling and Cleansing
- 3. Part of Speech Tagging
- 4. Parsing Structure in Text
- 5. NLP Applications
- Building your first NLP application
- Other NLP applications
- Summary
- 6. Text Classification
- 7. Web Crawling
- 8. Using NLTK with Other Python Libraries
- 9. Social Media Mining in Python
- 10. Text Mining at Scale
- 2. Module 2
- 1. Tokenizing Text and WordNet Basics
- Introduction
- Tokenizing text into sentences
- Tokenizing sentences into words
- Tokenizing sentences using regular expressions
- Training a sentence tokenizer
- Filtering stopwords in a tokenized sentence
- Looking up Synsets for a word in WordNet
- Looking up lemmas and synonyms in WordNet
- Calculating WordNet Synset similarity
- Discovering word collocations
- 2. Replacing and Correcting Words
- 3. Creating Custom Corpora
- Introduction
- Setting up a custom corpus
- Creating a wordlist corpus
- Creating a part-of-speech tagged word corpus
- Creating a chunked phrase corpus
- Creating a categorized text corpus
- Creating a categorized chunk corpus reader
- Lazy corpus loading
- Creating a custom corpus view
- Creating a MongoDB-backed corpus reader
- Corpus editing with file locking
- 4. Part-of-speech Tagging
- Introduction
- Default tagging
- Training a unigram part-of-speech tagger
- Combining taggers with backoff tagging
- Training and combining ngram taggers
- Creating a model of likely word tags
- Tagging with regular expressions
- Affix tagging
- Training a Brill tagger
- Training the TnT tagger
- Using WordNet for tagging
- Tagging proper names
- Classifier-based tagging
- Training a tagger with NLTK-Trainer
- 5. Extracting Chunks
- Introduction
- Chunking and chinking with regular expressions
- Merging and splitting chunks with regular expressions
- Expanding and removing chunks with regular expressions
- Partial parsing with regular expressions
- Training a tagger-based chunker
- Classification-based chunking
- Extracting named entities
- Extracting proper noun chunks
- Extracting location chunks
- Training a named entity chunker
- Training a chunker with NLTK-Trainer
- 6. Transforming Chunks and Trees
- Introduction
- Filtering insignificant words from a sentence
- Correcting verb forms
- Swapping verb phrases
- Swapping noun cardinals
- Swapping infinitive phrases
- Singularizing plural nouns
- Chaining chunk transformations
- Converting a chunk tree to text
- Flattening a deep tree
- Creating a shallow tree
- Converting tree labels
- 7. Text Classification
- Introduction
- Bag of words feature extraction
- Training a Naive Bayes classifier
- Training a decision tree classifier
- Training a maximum entropy classifier
- Training scikit-learn classifiers
- Measuring precision and recall of a classifier
- Calculating high information words
- Combining classifiers with voting
- Classifying with multiple binary classifiers
- Training a classifier with NLTK-Trainer
- 8. Distributed Processing and Handling Large Datasets
- Introduction
- Distributed tagging with execnet
- Distributed chunking with execnet
- Parallel list processing with execnet
- Storing a frequency distribution in Redis
- Storing a conditional frequency distribution in Redis
- Storing an ordered dictionary in Redis
- Distributed word scoring with Redis and execnet
- 9. Parsing Specific Data Types
- A. Penn Treebank Part-of-speech Tags
- 1. Tokenizing Text and WordNet Basics
- 3. Module 3
- 1. Working with Strings
- 2. Statistical Language Modeling
- Understanding word frequency
- Applying smoothing on the MLE model
- Develop a back-off mechanism for MLE
- Applying interpolation on data to get mix and match
- Evaluate a language model through perplexity
- Applying metropolis hastings in modeling languages
- Applying Gibbs sampling in language processing
- Summary
- 3. Morphology – Getting Our Feet Wet
- 4. Parts-of-Speech Tagging – Identifying Words
- 5. Parsing – Analyzing Training Data
- 6. Semantic Analysis – Meaning Matters
- 7. Sentiment Analysis – I Am Happy
- 8. Information Retrieval – Accessing Information
- 9. Discourse Analysis – Knowing Is Believing
- 10. Evaluation of NLP Systems – Analyzing Performance
- B. Bibliography
- Index
As you've seen in previous recipes, parse trees often have a variety of Tree label types that are not present in chunk trees. If you want to use parse trees to train a chunker, then you'll probably want to reduce this variety by converting some of these tree labels to more common label types.
First, we have to decide which Tree labels need to be converted. Let's take a look at that first Tree again:
Immediately, you can see that there are two alternative NP subtrees: NP-SBJ and NP-TMP. Let's convert both of those to NP. The mapping will be as follows:
|
Original Label |
New Label |
|---|---|
|
NP-SBJ |
NP |
|
NP-TMP |
NP |
In transforms.py is the function convert_tree_labels(). It takes two arguments: the Tree to convert and a label conversion mapping. It returns a new Tree with all matching labels replaced based on the values in the mapping:
from nltk.tree import Tree
def convert_tree_labels(tree, mapping):
children = []
for t in tree:
if isinstance(t, Tree):
children.append(convert_tree_labels(t, mapping))
else:
children.append(t)
label = mapping.get(tree.label(), tree.label())
return Tree(label, children)Using the mapping table we saw earlier, we can pass it in as a dict to convert_tree_labels() and convert the first parsed sentence from treebank:
>>> from transforms import convert_tree_labels
>>> mapping = {'NP-SBJ': 'NP', 'NP-TMP': 'NP'}
>>> convert_tree_labels(treebank.parsed_sents()[0], mapping)
Tree('S', [Tree('NP', [Tree('NP', [Tree('NNP', ['Pierre']), Tree('NNP', ['Vinken'])]), Tree(',', [',']), Tree('ADJP', [Tree('NP', [Tree('CD', ['61']), Tree('NNS', ['years'])]), Tree('JJ', ['old'])]), Tree(',', [','])]), Tree('VP', [Tree('MD', ['will']), Tree('VP', [Tree('VB', ['join']), Tree('NP', [Tree('DT', ['the']), Tree('NN', ['board'])]), Tree('PP-CLR', [Tree('IN', ['as']), Tree('NP', [Tree('DT', ['a']), Tree('JJ', ['nonexecutive']), Tree('NN', ['director'])])]), Tree('NP', [Tree('NNP', ['Nov.']), Tree('CD', ['29'])])])]), Tree('.', ['.'])])As you can see in the following diagram, the NP-* subtrees have been replaced with NP subtrees:
The convert_tree_labels() function recursively converts every child subtree using the mapping. The Tree is then rebuilt with the converted labels and children until the entire Tree has been converted.
The result is a brand new Tree instance with new subtrees whose labels have been converted.
The previous two recipes cover different methods of flattening a parse Tree, both of which can produce subtrees that may require mapping before using them to train a chunker. Chunker training is covered in the Training a tagger-based chunker recipe in Chapter 5, Extracting Chunks.
-
No Comment