Probabilistic Programming & Bayesian Methods for Hackers

by Cameron Davidson-Pilon

An intro to Bayesian methods and probabilistic programming from a computation/understanding-first, mathematics-second point of view.

http://camdavidsonpilon.github.io/Pro...

Genres: Computer Science, Programming, Science, Mathematics, Engineering, Technology, Nonfiction

134 pages, ebook

Reviews

[A Mig · Rating 4 out of 5]

My rating is for the nice tutorial that this 'book' is. It is not per se a book but a collection of python notebooks. For it to be rated as a book, concepts should be developed in more detail and the writing style changed to not look like a series of blog posts.

So - as a tutorial on Bayesian methods - it was very instructive with a wide variety of examples presented, nice illustrations, and obviously all the python codes. I will for sure come back to it later to redo some of the Bayesian predictions myself.

[Atti · Rating 5 out of 5]

An open collaborative and interactive book, what a great idea the author had in doing the book on github with Jupyter notebooks, all books should be like this one, i`m not sure how you would make money on them though.

So many great examples and ideas in this book, some were to advanced for me though so i have a lot of homework to do after reading this book.

The author was passionate he wanted to understand some of the solutions proposed to data science problems, like the dark matter solution on Kaggle so he got into scientific python and PyMC and this is his way of giving back to the community, he sets a good example for the rest of us.

[Ray · Rating 5 out of 5]

In my opinion, the strength of this text is that it doesn't present Bayesian methods as some arcane, specialized set of skills that few people ever master. Instead, they are shown as the sensible way to model and make decisions in an uncertain world. I especially liked the treatment of loss functions and maximizing payoffs rather than prediction accuracy. The chapter on priors was also quite good, though it did feel like a bit of a tour in a zoo without meaningful segues between sections. The Jupyter notebook format of the text is a great boon, and one which I would encourage all readers to take advantage of if they can.

[Elijah Tai · Rating 4 out of 5]

http://elijahtai.com/2017/11/28/refle...

[Andrei Barbu · Rating 4 out of 5]

Really a pymc tutorial. Read with students over the summer school. Lots of misinformation and many mistakes to be a real book about probabilistic programming.

[Evan Zamir · Rating 5 out of 5]

Really helpful book for getting started in Bayesian programming.

[THN · Rating 2 out of 5]

Not useful, for me.