Book/Website Recommendation

#1
I have a math background, but know very little statistics. I would love to find a book that shows one how to go from observational data to fitting a distribution to that data. I'm also Keen to learn R. So, if there is something that contains both subjects, all the better.

Thanks!
 

staassis

Active Member
#2
James, G., Witten, D., Hastie, T., & Tibshirani, R. (2017). An Introduction to Statistical Learning: with Applications in R (Corr. 7th printing). Springer New York.
 

ondansetron

TS Contributor
#3
It always makes me chuckle when people suggest running before someone knows how to walk... try starting with Blitzstein's Introduction to Probability found here. Literally a free course and text book to give you a decent undergraduate background in some basic elements needed to adequately approach statistical problems. After that, a book like the one noted above might be more appropriate.
 
#4
It always makes me chuckle when people suggest running before someone knows how to walk... try starting with Blitzstein's Introduction to Probability found here. Literally a free course and text book to give you a decent undergraduate background in some basic elements needed to adequately approach statistical problems. After that, a book like the one noted above might be more appropriate.
So, you’re saying one needs an undergraduate knowledge of probability in order to study statistics?
 

staassis

Active Member
#5
So, you’re saying one needs an undergraduate knowledge of probability in order to study statistics?
He is actually right. When you originally posted: "I have a math background", I thought that included probability theory. If not, you have to gain that knowledge before reading James et. al. If reading the book that @ondansetron suggested, focus on chapters 1-10. Chapters 11- 13 are important but can be skipped during the first reading.
 

ondansetron

TS Contributor
#6
He is actually right. When you originally posted: "I have a math background", I thought that included probability theory. If not, you have to gain that knowledge before reading James et. al. If reading the book that @ondansetron suggested, focus on chapters 1-10. Chapters 11- 13 are important but can be skipped during the first reading.
Agreed-- if you meant [calculus, linear algebra, geometry/trig, and some proofwork] as "math background" you definitely need the undergrad probability (and hopefully some mathematical statistics like Hogg and Craig or Wackerly/Mendenhall (at least) to get the most of ISL by Hastie et al. which is somewhat condensed overview of all of those topics in it). That being said, the ISL by Hastie is a good overview with some application.

The probability book I recommended isn't super rigorous, from a mathematical treatment, but has a great blend of logical thinking/problem solving with mathematical components as needed. It's not an easy course but it isn't the hardest/most dry.
 
#7
I guess I'm just looking for a introductory applied statistics book aimed at the undergrad level. Not looking to write a dissertation here. I do have a BS in Math/Computer Science, not including any statistics or probability. I took everything but.

Most of the books I'm looking at usually start a problem with "assume this data is normally distributed" or pick a distribution. I guess I'd just like to learn, at introductory level, how to get from data to distribution. Is that not possible? Do I really need to study at graduate level to do this? I know engineers and scientists use statistics without that rigorous of a background.

Thanks for all responses, but I'm not really up to self-studying Probability for 1+ year.
 

staassis

Active Member
#8
Thanks for all responses, but I'm not really up to self-studying Probability for 1+ year.
Well, if it takes you 1 year to go through that 1 undergraduate-level book then something is fundamentally wrong. You need to know the foundations of Probability Theory to study Statistics. Period.
 

ondansetron

TS Contributor
#9
I guess I'm just looking for a introductory applied statistics book aimed at the undergrad level. Not looking to write a dissertation here. I do have a BS in Math/Computer Science, not including any statistics or probability. I took everything but.

Most of the books I'm looking at usually start a problem with "assume this data is normally distributed" or pick a distribution. I guess I'd just like to learn, at introductory level, how to get from data to distribution. Is that not possible? Do I really need to study at graduate level to do this? I know engineers and scientists use statistics without that rigorous of a background.

Thanks for all responses, but I'm not really up to self-studying Probability for 1+ year.
I will clarify that none of these suggestions are graduate-level.

I think with your math background you can handle the probability book given. Mathematics is challenging. Statistics and probability are also hard subjects. Despite what you may see done, there is no quick fix for statistics. I know many engineers and scientists who use statistics without much formal background but most of them don't do it well (speaking from personal experience).

Doing statistics poorly can have many bad ramifications (speaking from a biomedical perspective, it can literally be life and death with patient research). Economically, this can costs millions of dollars if decisions are based on poorly done statistics; the goal you're mentioning would be very achievable with the background I suggested.

This may sound over the top but too often statistics is perceived as some set of cookbook calculations to be performed with software, but that couldn't be further from the truth.

I can't offer better recommendations than I already have done.