You're reading from Data Cleaning and Exploration with Machine Learning Get to grips with machine learning techniques to achieve sparkling-clean data quickly

Product type Paperback

Published in Aug 2022

Publisher Packt

ISBN-13 9781803241678

Length 542 pages

Edition 1st Edition

Concepts

Machine Learning

Author (1):

Michael Walker

View More author details

Table of Contents (23) Chapters

Preface

1. Section 1 – Data Cleaning and Machine Learning Algorithms

2. Chapter 1: Examining the Distribution of Features and Targets FREE CHAPTER

3. Chapter 2: Examining Bivariate and Multivariate Relationships between Features and Targets

4. Chapter 3: Identifying and Fixing Missing Values

5. Section 2 – Preprocessing, Feature Selection, and Sampling

6. Chapter 4: Encoding, Transforming, and Scaling Features

7. Chapter 5: Feature Selection

8. Chapter 6: Preparing for Model Evaluation

9. Section 3 – Modeling Continuous Targets with Supervised Learning

10. Chapter 7: Linear Regression Models

11. Chapter 8: Support Vector Regression

12. Chapter 9: K-Nearest Neighbors, Decision Tree, Random Forest, and Gradient Boosted Regression

13. Section 4 – Modeling Dichotomous and Multiclass Targets with Supervised Learning

14. Chapter 10: Logistic Regression

15. Chapter 11: Decision Trees and Random Forest Classification

16. Chapter 12: K-Nearest Neighbors for Classification

17. Chapter 13: Support Vector Machine Classification

18. Chapter 14: Naïve Bayes Classification

19. Section 5 – Clustering and Dimensionality Reduction with Unsupervised Learning

20. Chapter 15: Principal Component Analysis

21. Chapter 16: K-Means and DBSCAN Clustering

22. Other Books You May Enjoy

Feature scaling

Often, the features we want to use in our model are on very different scales. Put simply, the distance between the minimum and maximum values, or the range, varies substantially across possible features. For example, in the COVID-19 data, the total cases feature goes from 1 to almost 34 million, while aged 65 or older goes from 9 to 27 (the number represents the percentage of the population).

Having features on very different scales impacts many machine learning algorithms. For example, KNN models often use Euclidean distance, and features with greater ranges will have a greater influence on the model. Scaling can address this problem.

In this section, we will go over two popular approaches to scaling: min-max scaling and standard (or z-score) scaling. Min-max scaling replaces each value with its location in the range. More precisely, the following happens: