Ask HN: Bayesian Statistics – where do I start? | Hacker News Reader

Ask HN: Bayesian Statistics – where do I start?

11 pointscessor6 comments

A friend of mine mentioned that testing hypothesis in social sciences the "classical" way can be quite limiting and that Bayesian Statistics are "the future".

Could you explain why that is?

Do you have any recommendations on a good textbook?

View on HN

E. T. Jaynes' book "Probability Theory" is a tour de force and freely available[1]. The preface and first chapter answers your first question; it exceeds anything you'll probably read here, certainly from me.

[1] http://bayes.wustl.edu/etj/prob/book.pdf

Bayesian Reasoning and Machine Learning (BRML) by David Barber is an excellent introductory textbook. It's also free: http://www0.cs.ucl.ac.uk/staff/d.barber/brml/

Sounds like your friend is an acolyte of the LessWrong cult. That doesn't mean your friend is wrong. If you go on Lesswrong's website they have all sorts of articles on why Bayesian statistics is better.

I haven't looked into it very much, but I think they see Bayesian statistics as better because it better takes prior information into account.

Here's a persuasive essay on the topic by LessWrong's cult leader: http://lesswrong.com/lw/ul/my_bayesian_enlightenment/

free texts by Barber and MacKay: http://www.reddit.com/r/MachineLearning/comments/1jeawf/mach...

_________

course lecture notes: http://alumni.media.mit.edu/~tpminka/courses/10-602/schedule...

https://stat.duke.edu/~st118/sta250/

_____________

texts by Kruschke in R and Gelman/Carlin/Stern et al

http://www.stat.columbia.edu/~gelman/book/

http://www.r-bloggers.com/first-impressions-of-doing-bayesia...

I'm still learning about Bayesian Stats myself, but so far I've really enjoyed Think Bayes[0] by Allen Downey. He's a great writer and one of his other books was recently on the front page here.

0 - http://www.greenteapress.com/thinkbayes/

Thank you, I will give it a try, I enjoyed the think-x series for a while. You all have been helping me out in a great way, I will start here...

Reply on news.ycombinator.com