Python text mining

In order to be successful in this course, you will need to know how to program in Python. The expectation is that you have completed the first three courses in this Applied Data Science with Python series, specifically Course 1 on Introduction to Data Science in Python and Course 3 on Applied Machine Learning in Python, so that you are familiar with the numpy and pandas Python libraries for data manipulation, and scikit-learn toolkit for machine learning algorithms.

Primitive constructs in Text

• Sentences / input strings
• Words or Tokens
• Characters
• Document, larger files

In this course in python we are tolking about all this concepts and their properties

Week 1

Les liens utils:

1. https://docs.python.org/3/library/re.html

2. https://www.analyticsvidhya.com/blog/2014/11/text-data-cleaning-steps-python/

3. https://ieva.rocks/2016/08/07/cleaning-text-for-nlp/

4. https://chrisalbon.com/python/cleaning_text.html

Week 2

In this module, we will tolk to Natural Language

What is Natural Language ?

• Language used for everyday communication by humans
  • English
  • Chinese
  • spanish

compared to the artificial computer language

• Any computation, manipulation of natural language
• Natural language evolve
  • new words get added
  • old words lose popularity
  • language rules themselves may change.

NLT Task: A Broad Spectrum

• Computing words, counting frequency of words
• Finding sentence boundaries
• Part of speech tagging
• Parsing the sentence structure
• Identifying semantic roles
• Identifying entities in a sentences
• Finding which pronoun refers to which entity

An Introduction to NLTK

• NLTK: Natural Language Toolkit
• Open source library in Python
• Has support for most NLP tasks
• Also provides access to numerous text corpora

Usage of NLTK

• Importation
  import nltk

• Let's get some text corpora
  nltk.download()

  from nltk.dowload()
  

  for more information see lab week2

Tokenization

• Recall splitting a sentence into words / tokens

Part-of-speech (POS) Tagging

• Recall high school grammar: nouns, verbs, adjectives,...
  

Ambiguity in POS Tagging

Parsing Sentence Structure

Ambiguity in Parsing

POS Tagging & Parsing Complexity

Task Home Concepts

Examples of Text Classification

Supervised Learning

Supervised Classification Step

Supervised Classification Model

Divide Dataset in two parts

Classification paradigms

Questions to ask in Supervised Learning

Why is textual data unique ?

Types of textual features (1)

Types of textual features (2)

Types of textual features (3)

Naive Bayes Classifiers

Case study: Classifying text search queries

Probabilistic model

Bayes' Rule

Naive Bayes Classification

Example classification

Naïve Bayes: Learning parameters

Naïve Bayes: Smoothing

Take Home Concepts

Two Naïve Bayes Variants For Text

Support Vector Machine

Decision Boundaries

Choosing a Decision Boundary

Finding a Linear Boundary

SVM: Multi-class classification

SVM Parameters (1): Parameter C

SVM Parameters (2): Others Params

Take Home Messages

Using Sklearn's NaiveBayesClassifier

Using Sklearn's SVM Classifier

Model Selection in Scikit-learn

Supervised Text Classification in NLTK

Using NLTK's NaiveBayesClassifier

Using NLTK's SkearnClassifier

Take Home Concept

Application of semantic similarity

WordNet

Semantic similarity using WordNet

Coming back to our deer example

Similarity with NLP in Python

Distributional Similarity: Context

Strength of association between words

Take Home Concepts

What is Topic Modeling ?

Generative Models for Text

Generative Model can be complex

Latent Dirichlet Allocation (LDA)

Topic Modeling in Practice

Topic Modeling: Summary

Working with LDA in Python

Take Home Concepts

Information is hidden in free-text

Information Extraction

Fields of Interest

Named Entity Recognition

Approche to identify named entities

Relation extraction

Co-reference resolution

Question Answering

Take Home Concepts

Additional Resources & Readings

• http://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf

• https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation

• https://en.wikipedia.org/wiki/Plate_notation

• https://www.nltk.org/howto/wordnet.html