5
$\begingroup$

I posted a question earlier where I mentioned that I am interested in learning Machine Learning but that my background in statistics and probability is pretty weak.

Recently I previewed pages of 2 books which seem to quite suit my requirements. I was just wanting to know what the community thinks about my possible choices, given my background and goals.

(1) All of Statistics: A Concise Course in Statistical Inference by Larry Wasserman. Looks nice to me but the author does not provide the answers to exercise problems (let alone a solution manual).

(2) A seemingly less well known but again a concise book: Probability and Statistics for Computer Scientists by Michael Baron

Any thoughts on these books or any good alternatives? Just to emphasize, I am a CS student looking to get to speed with probability and statistics for Machine Learning. I am at this point only looking for books suggestions not websites or videos.

Improve this question
$\endgroup$
3
  • 1
    $\begingroup$ "Looks nice to me but the author does not provide the answers to exercise problems (let alone a solution manual)." As a general rule, solutions manuals are rarely sold to the general public when a book is intended as a textbook, but are provided by the publisher to instructors who adopt the book as the text for the course. Some books include answers (not necessarily solutions) to the exercises, or to the odd-numbered exercises so that the student can check his/her work for self-learning while the even-numbered exercises can be assigned as homework by the instructor. $\endgroup$
    – Dilip Sarwate
    Commented Dec 5, 2011 at 1:50
  • 1
    $\begingroup$ @DilipSarwate : Yes, I am aware of that. It's just that it's a problem if none of the questions have answers, especially if you are going to be using the book for self study. $\endgroup$
    – user721975
    Commented Dec 5, 2011 at 2:12
  • 3
    $\begingroup$ All of Statistics is definitely the book! As the author of several stat books, I am reluctant to put solution manuals on line for two reasons: (a) this reduces the appeal of the book for instructors who want to use the exercises from the book and (b) this induces an endless flow of emails from readers who do not understand (or disagree with) the solutions. However, we did put solutions for Bayesian Core and Introducing Monte Carlo Methods with R on arXiv. $\endgroup$
    – Xi'an
    Commented Feb 4, 2012 at 9:10

4 Answers 4

Reset to default
5
$\begingroup$

I have the 8th edition of Modern Elementary Statistics which I see has a companion answers book to odd numbered questions. I also have an older version of Ott's Introduction to Statistical Methods and Data Analysis which I find incredibly useful to pointing out to colleagues about how I would love to see GLM models specified in journal articles. When I was studying, I found Harraway's book incredibly useful but sadly it never went to another edition. He's a really good writer that had good social science examples.

Improve this answer
$\endgroup$
3
$\begingroup$

The first book "All of Statistics" is nice book to read. I dont know about second book. However, I would recommend the following book. It is really nice book to get an idea whats going on in statistics and probability.

A Modern Introduction to Probability and Statistics: Understanding Why and How (Springer Texts in Statistics)

by: F.M. Dekking, C. Kraaikamp, H.P. Lopuhaä, L.E. Meester

Improve this answer
$\endgroup$
2
$\begingroup$

I still like DeGroot / Schervish. Please, take a look:

http://www.amazon.com/Probability-Statistics-Edition-Morris-DeGroot/dp/0321500466

It is highly self-contained, starting with basic set theory, covering all the standard material on probability (without measure theory). After that, the inference part shines, with excellent explanations of sufficiency, estimation, the Neyman-Pearson lemma, hypothesis testing. He goes through nonparametric methods and ANOVA at the end of the book. DeGroot writing is not fast paced and is extremely clear. Since he was a Bayesian statistician (as is Schervish), the discussion of Bayesian topics is more detailed than we usually see in other books at the same level. The third edition covers simulation at the end, including MCMC tools.

Improve this answer
$\endgroup$
0
1
$\begingroup$

I worked through the entire book about a year ago (2019) and put most of my solutions to Chapters 8–16 on github.

Please comment and correct me if I’m wrong, but I believe the free pdf released is a preprint, and the printed book is the official version. I used the free pdf (and included a copy of it in my git repo) but there will be some differences (I think some chapters are in a different order, some problems are different, etc), so keep that in mind if you decide to follow the physical text.

In my git repo I include a copy of the free pdf, and have some links pointing to old class versions that Larry taught at CMU. The homeworks, and included solutions, follow the text pretty closely, so they are helpful to look at when working through the text.

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ Which book, please? The OP lists two possible. Would you please also provide an appropriate link to the github? $\endgroup$
    – Carl
    Commented Feb 27, 2020 at 3:18
  • $\begingroup$ Does CMU stand for Carnegie Mellon Univ ? If so, then it instantly leads me to the classical textbook written by the CMU professors : Probability and Statistics by Morris H. DeGroot and Mark J. Schervish. It's among the lists I want to read, although probably not the main one; I will go to that book if I need a complete or better explanation on some particular topics, such as Moment Generating Functions. $\endgroup$
    – Lex Soft
    Commented Jul 6, 2022 at 3:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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