CODEWITHSUNDEEP
machine-learningbayesian
suemi
suemi

Reputation: 69

a summary of frequentist view in machine learning

I have read ESL and PRML,the PRML is easy for me to understand, as well with the Bayesian view.But the PRML is too difficult for people without statistic background like me .Is there anybody able to give a summary of frequentist view or explain it briefly?

Upvotes: 1

Views: 744

Answers (1)

rano
rano

Reputation: 5666

It is a bit hard, in my opinion, to make it brief. It's like banalizing two different visions of a world.

Nevertheless, a very short reduction to a single argument can be this:

  • frequentists do not treat hidden parameters as random variables and do not seek to describe them by subjective priors.

Here is a good discussion that goes in depth touching arguments as parameter estimation algorithmic complexty in the two approaches.

This is an old discussion, arisen in the famous Langford's blog.

And, to confirm my point about the requirment of a bit of statistics at least to understand the different perspectives, here is a video.

Upvotes: 2

Related Questions