Book Image

Python Machine Learning, Second Edition - Second Edition

By : Sebastian Raschka, Vahid Mirjalili
Book Image

Python Machine Learning, Second Edition - Second Edition

By: Sebastian Raschka, Vahid Mirjalili

Overview of this book

Publisher's Note: This edition from 2017 is outdated and is not compatible with TensorFlow 2 or any of the most recent updates to Python libraries. A new third edition, updated for 2020 and featuring TensorFlow 2 and the latest in scikit-learn, reinforcement learning, and GANs, has now been published. Machine learning is eating the software world, and now deep learning is extending machine learning. Understand and work at the cutting edge of machine learning, neural networks, and deep learning with this second edition of Sebastian Raschka’s bestselling book, Python Machine Learning. Using Python's open source libraries, this book offers the practical knowledge and techniques you need to create and contribute to machine learning, deep learning, and modern data analysis. Fully extended and modernized, Python Machine Learning Second Edition now includes the popular TensorFlow 1.x deep learning library. The scikit-learn code has also been fully updated to v0.18.1 to include improvements and additions to this versatile machine learning library. Sebastian Raschka and Vahid Mirjalili’s unique insight and expertise introduce you to machine learning and deep learning algorithms from scratch, and show you how to apply them to practical industry challenges using realistic and interesting examples. By the end of the book, you’ll be ready to meet the new data analysis opportunities. If you’ve read the first edition of this book, you’ll be delighted to find a balance of classical ideas and modern insights into machine learning. Every chapter has been critically updated, and there are new chapters on key technologies. You’ll be able to learn and work with TensorFlow 1.x more deeply than ever before, and get essential coverage of the Keras neural network library, along with updates to scikit-learn 0.18.1.
Table of Contents (18 chapters)
17
Index

Maximum margin classification with support vector machines


Another powerful and widely used learning algorithm is the Support Vector Machine (SVM), which can be considered an extension of the perceptron. Using the perceptron algorithm, we minimized misclassification errors. However, in SVMs our optimization objective is to maximize the margin. The margin is defined as the distance between the separating hyperplane (decision boundary) and the training samples that are closest to this hyperplane, which are the so-called support vectors. This is illustrated in the following figure:

Maximum margin intuition

The rationale behind having decision boundaries with large margins is that they tend to have a lower generalization error whereas models with small margins are more prone to overfitting. To get an idea of the margin maximization, let's take a closer look at those positive and negative hyperplanes that are parallel to the decision boundary, which can be expressed as follows:

If we subtract those...