This is probably the best book for beginners who want to get into ML. This provides a comprehensive introduction to all the popular ML algorithms and also teaches the popular libraries of python which is used for analysis, namely pandas. The book is math-heavy but does not make the person feel overwhelmed with the amount of statistics present (the equations appear scary but once you read the descriptions it becomes easy). There are ready-made code templates printed throughout the book which can be plugged and played in other algorithms. For freshers it is suggested to type out the algorithm as it helps in learning the algorithm and trouble shooting the same.
Decent book. I will suggest you use this book only if you are enrolled in some ML course, explanations are not too detailed, so it may not work for self study.
Covers many techniques including forecasting models and text analytics which are usually omitted from ML books. Doesn't covers SVMs though.
This book is actually a collection of tutorials, so not comparable to Bishop/Duda etc in rigor but complements them as you get to see how to implement things.