3
$\begingroup$

I'm a Scientific Computing PhD student with a strong background in linear algebra, calculus, and numerical methods, and some exposure to probability theory. I've taken a grad level machine learning class, and I'd say I understood the material pretty well. I don't really have a stats & probability background aside from the grad courses I've taken that used probability theory, and undergrad stats and AP stats.

Anyways, I just started reading ESL yesterday, and I must admit that I'm having a lot of difficulty following. Are there any supplemental material to go along with this book?

For example, in Chapter 3, everything made sense until Eq. (3.8). Everything after that in that section is going over my head. When they wrote "The $N − p − 1$ rather than $N$ in the denominator makes $\hat{\sigma}^2$ an unbiased estimate of $\sigma^2$," I was like "OK I don't know where this came from, but I'll just take author's word for it." And then I proceeded to the following paragraph, and got stuck again, this time on "We now assume that (3.1) is the correct model for the mean; that is, the conditional expectation of Y is linear in X1, . . . , Xp. We also assume that the deviations of Y around its expectation are additive and Gaussian. Hence..."

If this is going to be a common occurrence throughout the book, then I think I'd rather just stop reading it and read up on ML methods online.

I'm reading this book for two purposes: (1) preparing for quant interviews (2) to learn ML in more depth

Improve this question
$\endgroup$
5
  • 4
    $\begingroup$ Sounds like you are after a basic book on regression theory. $\endgroup$
    – Glen_b
    May 18 '20 at 3:45
  • $\begingroup$ @Glen_b-ReinstateMonica No, because I'm interested in other methods as well, not just regression. $\endgroup$
    – 24n8
    May 18 '20 at 3:48
  • 4
    $\begingroup$ It looks like that's where you need to begin. $\endgroup$
    – Glen_b
    May 18 '20 at 4:13
  • 3
    $\begingroup$ Right: one doesn't just jump right in learning everything at once. A big part of the foundation you seek is familiarity with the language, terminology, concepts, and important questions in the field. No amount of mathematical knowledge prepares you for that, so you shouldn't view studying elementary material like regression as being superfluous, overly narrow, or beneath you: it will indeed make reading ESL much easier. cc @Glen_b $\endgroup$
    – whuber ♦
    May 18 '20 at 13:23
  • $\begingroup$ Please see a solution manual here yuhangzhou88.github.io/ESL_Solution $\endgroup$
    – Yuhang
    Apr 17 at 21:57
1
$\begingroup$

Somewhat following @whuber and @Glen_b comments, maybe worth checking L. Wasserman All of Statistics. Besides, Springer has recently made the pdf free to download.

A couple of quotes from the preface:

This book is for people who want to learn probability and statistics quickly. It is suitable for graduate or advanced undergraduate students in computer science, mathematics, statistics, and related disciplines. [...]

Using fancy tools like neural nets, boosting, and support vector machines without understanding basic statistics is like doing brain surgery before knowing how to use a band-aid.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.