The applications of Machine Learning are endless. “Practical Machine Learning” is a resource intended for programmers with no background in Machine Learning yet are interested in building a solid foundation to adequately implement real-world Machine Learning Solutions. No background in Linear Algebra, Probability, or Calculus is assumed. Sufficient Machine Learning theory is covered where required; The keen readers are also referred to external resources where needed. The only prerequisite is a good programming background preferably in Python. The primary libraries used are Numpy and Scikit-Learn.

Why Did the Author Write this Book

Programmers and developers- with no background in Machine Learning- often resort to searching for quick pre-written solutions by taking Machine Learning code from disparate sources without adequately understanding the code. This approach results in inelegant solutions which don’t sufficiently solve the real-world problems. Moreover, bugs can easily be introduced in solutions when placed in the new environment, and debugging Machine Learning requires satisfactory knowledge of Machine Learning: one needs enough understanding of how things are working beneath the hood. This book aims to fill that gap i.e. it aims to build the foundation necessary to write good Machine Learning solutions. For example, where educational, we will see how Scikit-Learn is working under the hood.

Learning By Doing

“The best way of learning about anything is by doing”. This could not be more true for learning to implement Machine Learning solutions. Exercises (in the form of 10 Tests) are the backbone of this book. They’ll make you think deeply about Machine Learning. ‘Tests’ is a misnomer; They are primarily there to teach you- not to test you. Each test has two versions: one with the solution and one without it. The version with the solution is meant to be used guardedly since reading the solutions will defeat the purpose if you do not challenge yourself first. Facial Recognition with dimensionality reduction (Test 6), image compression using Clustering (Test 5), stocks price prediction (Test 3), predicting political inclination of MPs of UK based on their tweets (Test 9), and recognizing a digit from its image (Test 4) are a few of the great applications of Machine Learning the reader will get hands-on experience with. Thereby, building a strong Machine Learning base. The exercises are created keeping in mind readers with diverse backgrounds. Even if you’ve years of experience in Machine Learning, the odds are the exercises will challenge you and make you ponder on a different level.

Is This Book About Deep Learning

Though Deep Learning is briefly discussed and used in the Natural Language Processing module, Deep Learning is not the focus of this book.

About the Author

The author is a freelance Machine Learning, ML Ops, and data engineer who- as an Artificial Intelligence expert- has been providing services to companies of various sizes all over the world.


This online version of the book will remain available for free.
If you notice any typos or have any suggestions, feel free to open a pull request here.