Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI

by Mark Ryan, Luca Massaron

504 pages, Paperback

Published March 25, 2025

About the author

Librarian Note: There is more than one author by this name in the Goodreads data base.

Reviews

[Tomáš Šipka · Rating 5 out of 5]

This book is an excellent learning resource for machine/deep learning on tabular data. It guides you through the entire process—from data analysis and model selection to deployment. The book is packed with Python examples, allowing you to get hands-on experience right away.

A prerequisite for this book is knowledge of Python, but since it’s almost a must-have skill in the ML community these days, I don’t see this as a drawback. On the other hand, it does not require strong prior knowledge of machine learning, as all the algorithms are clearly explained. You don’t even need a mathematical background, since the concepts are presented without complex equations. Depending on your background, this might be seen as either an advantage or a limitation.

However, if you are interested in diving deeper, the book provides plenty of references to research papers on various algorithms. In my opinion, this makes it a great resource for absolute beginners in ML as well as experienced engineers looking for a solid starting point (and more) in working with tabular data.

I particularly appreciate that this book shares knowledge on a less common but nonetheless important topic compared to computer vision, NLP, generative AI, etc.

[Ninoslav Cerkez · Rating 5 out of 5]

I really like this book. With so many machine learning books available today, it can be difficult to choose truly valuable titles worthy of investing time and money—but this one certainly is. It focuses specifically on tabular data, a format that remains predominant in many ML cases. The authors thoroughly explore the most effective classical machine learning algorithms and complement this coverage with deep learning approaches. Moreover, they don't stop at model creation; they also discuss deploying these ML models into production and building pipelines. In short, this book provides comprehensive coverage of machine learning for tabular data.

[Giovanni · Rating 5 out of 5]

When it comes to a number of fields, tabular data is extremely relevant: from post impressions to sensor readings...a lot of today's information is put into "tables" and being able to extract value from them is extremely important.

This book consists in an extensive overview covering both the "good, old methods" and the most recent techniques.

A must read for anyone who is interested into extracting more value and information from their tabular data!