Elevate your natural language processing skills with Gensim and become proficient in handling a wide range of NLP tasks and projects
Key Features
- Advance your NLP skills with this comprehensive guide covering detailed explanations and code practices
- Build real-world topical modeling pipelines and fine-tune hyperparameters to deliver optimal results
- Adhere to the real-world industrial applications of topic modeling in medical, legal, and other fields
- Purchase of the print or Kindle book includes a free PDF eBook
Book Description
Navigating the terrain of NLP research and applying it practically can be a formidable task made easy with The Handbook of NLP with Gensim. This book demystifies NLP and equips you with hands-on strategies spanning healthcare, e-commerce, finance, and more to enable you to leverage Gensim in real-world scenarios.
You’ll begin by exploring motives and techniques for extracting text information like bag-of-words, TF-IDF, and word embeddings. This book will then guide you on topic modeling using methods such as Latent Semantic Analysis (LSA) for dimensionality reduction and discovering latent semantic relationships in text data, Latent Dirichlet Allocation (LDA) for probabilistic topic modeling, and Ensemble LDA to enhance topic modeling stability and accuracy.
Next, you’ll learn text summarization techniques with Word2Vec and Doc2Vec to build the modeling pipeline and optimize models using hyperparameters. As you get acquainted with practical applications in various industries, this book will inspire you to design innovative projects. Alongside topic modeling, you’ll also explore named entity handling and NER tools, modeling procedures, and tools for effective topic modeling applications.
By the end of this book, you’ll have mastered the techniques essential to create applications with Gensim and integrate NLP into your business processes.
What you will learn
- Convert text into numerical values such as bag-of-word, TF-IDF, and word embedding
- Use various NLP techniques with Gensim, including Word2Vec, Doc2Vec, LSA, FastText, LDA, and Ensemble LDA
- Build topical modeling pipelines and visualize the results of topic models
- Implement text summarization for legal, clinical, or other documents
- Apply core NLP techniques in healthcare, finance, and e-commerce
- Create efficient chatbots by harnessing Gensim's NLP capabilities
Who this book is for
This book is for data scientists and professionals who want to become proficient in topic modeling with Gensim. NLP practitioners can use this book as a code reference, while students or those considering a career transition will find this a valuable resource for advancing in the field of NLP. This book contains real-world applications for biomedical, healthcare, legal, and operations, making it a helpful guide for project managers designing their own topic modeling applications.
Author(s): Chris Kuo
Publisher: Packt Publishing
Year: 2023
Language: English
Pages: 310
The Handbook of NLP with Gensim
Contributors
About the author
About the reviewers
Preface
Why read this book?
What is Gensim
Who this book is for
What this book covers
To get the most out of this book
Download the example code files
Data for this book
Conventions used
Get in touch
Share Your Thoughts
Download a free PDF copy of this book
Part 1: NLP Basics
Chapter 1: Introduction to NLP
Introduction to natural language processing
NLU + NLG = NLP
NLU
NLG
Gensim and its NLP modeling techniques
BoW and TF-IDF
LSA/LSI
Word2Vec
Doc2Vec
LDA
Ensemble LDA
Topic modeling with BERTopic
Common NLP Python modules included in this book
spaCy
NLTK
Summary
Questions
References
Chapter 2: Text Representation
Technical requirements
What word embedding is
Simple encoding methods
One-hot encoding
BoW
Bag-of-N-grams
What TF-IDF is
Shining applications of BoW and TF-IDF
Coding – BoW
Gensim for BoW
scikit-learn for BoW (CountVectorizer)
Coding – Bag-of-N-grams
Gensim for N-grams
scikit-learn for N-grams
NLTK for N-grams
Coding – TF-IDF
Gensim for TF-IDF
scikit-learn for TF-IDF
Summary
Questions
References
Chapter 3: Text Wrangling and Preprocessing
Technical requirements
Key steps in NLP preprocessing
Tokenization
Lowercase conversion
Stop word removal
Punctuation removal
Stemming
Lemmatization
Coding with spaCy
spaCy for lemmatization
spaCy for PoS
Coding with NLTK
NLTK for tokenization
NLTK for stop-word removal
NLTK for lemmatization
Coding with Gensim
Gensim for preprocessing
Gensim for stop-word removal
Gensim for stemming
Building a pipeline with spaCy
Summary
Questions
References
Part 2: Latent Semantic Analysis/Latent Semantic Indexing
Chapter 4: Latent Semantic Analysis with scikit-learn
Technical requirements
Understanding matrix operations
An orthogonal matrix
The determinant of a matrix
Understanding a transformation matrix
A transformation matrix in daily life examples
Understanding eigenvectors and eigenvalues
An introduction to SVD
Truncated SVD
Truncated SVD for LSI
Coding truncatedSVD with scikit-learn
Using TruncatedSVD
randomized_SVD
Using TruncatedSVD for LSI with real data
Loading the data
Creating TF-IDF
Using TruncatedSVD to build a model
Interpreting the outcome
Summary
Questions
Chapter 5: Cosine Similarity
Technical requirements
What is cosine similarity?
How cosine similarity is used in images
How to compute cosine similarity with scikit-learn
Summary
Questions
References
Chapter 6: Latent Semantic Indexing with Gensim
Technical requirements
Performing text preprocessing
Performing word embedding with BoW and TF-IDF
BoW
TF-IDF
Modeling with Gensim
BoW
TF-IDF
Using the coherence score to find the optimal number of topics
Saving the model for production
Using the model as an information retrieval tool
Loading the dictionary list
Preprocessing the new document
Scoring the document to get the latent topic scores
Calculating the similarity scores with the new document
Finding documents with high similarity scores
Summary
Questions
References
Part 3: Word2Vec and Doc2Vec
Chapter 7: Using Word2Vec
Technical requirements
Introduction to Word2Vec
Advantages of Word2Vec
Reviewing the real-world applications of Word2Vec
Introduction to Skip-Gram (SG)
Data preparation
The input and output layers
The hidden layer
Should I remove stop words for training Word2Vec?
Model computation
Introduction to CBOW
Using a pretrained model for semantic search
Adding and subtracting words/concepts
Example 1
Example 2
Visualizing Word2Vec with TensorBoard
Training your own Word2Vec model in CBOW and Skip-Gram
Load the data
Text preprocessing
Training your own Word2Vec model in CBOW
Training your own Word2Vec model in Skip-Gram
Visualizing your Word2Vec model with t-SNE
Comparing Word2Vec with Doc2Vec, GloVe, and fastText
Word2Vec versus Doc2Vec
Word2Vec versus GloVe
Word2Vec versus FastText
Summary
Questions
References
Chapter 8: Doc2Vec with Gensim
Technical requirements
From Word2Vec to Doc2Vec
PV-DBOW
The input layer
The hidden layer
The output layer
Model optimization
PV-DM
The real-world applications of Doc2Vec
Doc2Vec modeling with Gensim
Text preprocessing for Doc2Vec
Modeling
Saving the model
Saving the training data
Putting the model into production
Loading the model
Loading the training data
Use case 1 – find similar articles
Use case 2 – find relevant documents based on keywords
Tips on building a good Doc2Vec model
Summary
Questions
References
Part 4: Topic Modeling with Latent Dirichlet Allocation
Chapter 9: Understanding Discrete Distributions
Technical requirements
The basics of discrete probability distributions
Bernoulli distributions
The formal definition of a Bernoulli distribution
What does it look like?
Fun facts
Binomial distributions
The real-world examples
The formal definition of a binomial distribution
What does it look like?
Plotting it with Python
Fun facts
Multinomial distributions
The real-world examples
The formal definition of a multinomial distribution
What does it look like?
Fun facts
Beta distributions
The real-world examples
The formal definition of a beta distribution
What does it look like?
The beta distribution in Bayesian inference
Fun fact
Dirichlet distributions
Real-world examples
The formal definition of a Dirichlet distribution
What is a simplex?
What does the Dirichlet distribution look like?
The Dirichlet distribution in Bayesian inference
Fun fact
Summary
Questions
References
Chapter 10: Latent Dirichlet Allocation
What is generative modeling?
Discriminative modeling
Generative modeling
Bayes’ theorem
Expectation-Maximization (EM)
Understanding the idea behind LDA
Dirichlet distribution of topics
Understanding the structure of LDA
Variational inference
Variational E-M
Gibbs sampling in LDA
Variational E-M versus Gibbs sampling
Summary
Questions
References
Chapter 11: LDA Modeling
Technical requirements
Text preprocessing
Preprocessing
Experimenting with LDA modeling
A model built on BoW data
A model built on TF-IDF data
Building LDA models with a different number of topics
Models built on BoW data
Models built on TF-IDF data
Determining the optimal number of topics
Using the model to score new documents
Text preprocessing
Scoring new texts
Outcome
Summary
Questions
References
Chapter 12: LDA Visualization
Technical requirements
Designing an infographic
Data visualization with pyLDAvis
The interactive graph
Summary
Questions
References
Chapter 13: The Ensemble LDA for Model Stability
Technical requirements
From LDA to Ensemble LDA
The process of Ensemble LDA
Understanding DBSCAN and CBDBSCAN
DBSCAN
CBDBSCAN (Checkback DBSCAN)
Building an Ensemble LDA model with Gensim
Preprocessing the training data
Creating text representation with BOW and TF-IDF
Saving the dictionary
Building the Ensemble LDA model
Scoring new documents
Summary
Questions
References
Part 5: Comparison and Applications
Chapter 14: LDA and BERTopic
Technical requirements
Understanding the Transformer model
Understanding BERT
Describing how BERTopic works
BERT – word embeddings
UMAP – reduce the dimensionality of embeddings
HDBSCAN – cluster documents
c-TFIDF – create a topic representation
Maximal Marginal Relevance
Building a BERTopic model
Loading the data – no text preprocessing
Modeling
Reviewing the results of BERTopic
Getting the topic information
Inspecting the keywords of a single topic
Getting document information
Getting representative documents
Visualizing the BERTopic model
Visualizing topics
Visualizing the hierarchy of topics
Visualizing the top words of topics
Visualizing on a heatmap
Predicting new documents
Using the modular property of BERTopic
Word embeddings
Dimensionality reduction
Clustering
Comparing BERTopic with LDA
Approach
Word embeddings
Text preprocessing
Language understanding
Topic clarity
Determination of the number of topics
Determination of word significance in a topic
Summary
Questions
References
Chapter 15: Real-World Use Cases
Word2Vec for medical fraud detection
Background
Questions
NLP solution
Takeaways
Background
Questions
NLP solution
Takeaways
Background
Questions
NLP solution
Takeaways
Comparing LDA/NMF/BERTopic on Twitter/X posts
Background
Questions
NLP solution
Takeaways
Interpretable text classification from electronic health records
Background
Questions
NLP solution
Takeaways
BERTopic for legal documents
Background
Questions
NLP solution
Takeaways
Word2Vec for 10-K financial documents to the SEC
Background
Questions
NLP solution
Takeaways
Summary
References
Assessments
Chapter 1 – Introduction to NLP
Chapter 2 – Text Representation
Chapter 3 – Text Wrangling and Preprocessing
Chapter 4 – Latent Semantic Analysis with scikit-learn
Chapter 5 – Cosine Similarity
Chapter 6 – Latent Semantic Indexing with Gensim
Chapter 7 – Using Word2Vec
Chapter 8 – Doc2Vec with Gensim
Chapter 9 – Understanding Discrete Distributions
Chapter 10 – Latent Dirichlet Allocation
Chapter 11 – LDA Modeling
Chapter 12 – LDA Visualization
Chapter 13 – The Ensemble LDA for Model Stability
Chapter 14 – LDA and BERTopic
Index
Why subscribe?
Other Books You May Enjoy
Packt is searching for authors like you
Share Your Thoughts
Download a free PDF copy of this book
