You're reading from Scikit-learn Cookbook Over 80 recipes for machine learning in Python with scikit-learn

Product type Paperback

Published in Dec 2025

Last Updated in Sep 2025

Publisher Packt

ISBN-13 9781836644453

Length 414 pages

Edition 3rd Edition

Languages

Python

Tools

Scikit-learn

Concepts

Machine Learning

Author (1):

John Sukup

View More author details

Table of Contents (14) Chapters

1. Scikit-learn Cookbook, Third Edition: Over 80 recipes for machine learning in Python with scikit-learn

2. Chapter 1: Common Conventions and API Elements of scikit-learn FREE CHAPTER

3. Chapter 2: Pre-Model Workflow and Data Preprocessing

4. Chapter 3: Dimensionality Reduction Techniques

5. Chapter 4: Building Models with Distance Metrics and Nearest Neighbors

6. Chapter 5: Linear Models and Regularization

7. Chapter 6: Advanced Logistic Regression and Extensions

8. Chapter 7: Support Vector Machines and Kernel Methods

9. Chapter 8: Tree-Based Algorithms and Ensemble Methods

10. Chapter 9: Text Processing and Multiclass Classification

11. Chapter 10: Clustering Techniques

12. Chapter 11: Novelty and Outlier Detection

13. Chapter 12: Cross-Validation and Model Evaluation Techniques

14. Chapter 13: Deploying scikit-learn Models in Production

Text Vectorization Techniques

Text vectorization techniques transform raw text into numerical representations, enabling ML algorithms to effectively analyze textual data. Techniques such as Bag of Words (BoW), TF-IDF, and word embeddings offer different approaches for capturing textual features and semantic meaning. Keep in mind, however, that even before text vectorization begins, a computer still must have a way to represent a letter in the first place. This is typically done using ASCII or Unicode encodings as an initial way to represent and store text in computer memory. This recipe will provide examples of different methodologies for text vectorization.

Getting ready

We'll load necessary libraries and prepare textual data for vectorization. This time we will use the reuters corpus from NLTK. It contains 10,788 news documents and approximately 1.3 million words with over 90 possible topics for classification. We will just classify based on the first category for simplicity.

...