@@TranquilSeaOfMath LibGen is a way to get books for free, he's saying that by fuelling people's interest in the book, he indirectly pushes them to pirate the book more
This is a great book. My PhD program's comprehensive exam's mathematical stats part is mostly based on this book. I was (forced) to read it cover to cover since I am not from a stats undergrad program, but I have to say it's fun to read and I did every exercise and it is a lot of fun. I have graduated for a few years but I am still keeping a copy with me.
Many years ago I was failing second year stats (no need to mention the text...) until I discovered the "Honours textbook" Casella & Burger. Now it is one of the most valuable books I own. Seriously, I cannot praise it highly enough.
I loved this book back in my CS grad school days. I took a few stats courses outside the dept. and one of the courses used this book and I fell in love with stats primarily due to this book. I still have a copy that I go back to once in a while.
I did not expect to be buying a book on mathematical statistics but here I am. I feel like I could learn a lot from it as a compsci graduate. We really did not learn enough maths.
Personally I really like the various Statistical Learning books by Hastie/Tibsharini - building intuition is probably the most important thing you need in statistics and they were absolutely incredible in doing that (rather than just focusing on dry proofs where you may understand how something is true but not any of the insights behind this). Probably not a good first text as it's more geared towards specifically the applications of statistics but probably the best textbook I've ever read
Thank you so much - for the advice *and also* for keeping the info short and compressed. I wish more RU-vidrs and podcasters would follow this method in knowledge heavy topics.
I think In my math stats class we used a different book but my teacher had a print out of that diagram of the distributions from this book that my instructor gave us
“The first two chapters should be enough to get you into a data science position.” Yeah, only how do I convey this information to the hiring major, whose background is in communications.
"The first two chapters are sufficient for an entry level Data Science position." This is a really silly statement. No. Those first two chapters are just about probability and cover no statistics at all, which is the most important thing for a Data Scientist. If you don't know what a hypothesis test is, you're not ready to be a Data Scientist.
The good news about Statitics is that you don't need much probability theory to already have the tools to dive into statistics. Measure theory is a bit of a waste of time in Statistics, linear algebra is way more important.
Hm, the first two chapters say nothing about inference and maybe statistics though. I wouldn't have thought that that's enough for a job in data science. But it is what it is. Thanks for sharing.
@@xvzf8147 Wait, wait, wait. I don't have any experience in Data Science jobs, I''m just making an observation. If you say it's enough, then I'd rather trust your opinion than my initial imrpession.
@@PedroTricking intro to Statistical Learning covers the methods you would use in a work setting. This is books is theoretical stats. Most jobs do not require deep theoretical knowledge. That said, I would say that this helps a lot.
There's another reaaaaaaally good book going into the coding part that is great for those interested in Data Science or generally want to excel as a Statistician. "Computer Age Statistical Inference" is what it is called and you can find a free copy online courtesy of one of the universities which one of the author's actually worked at. I'm reading that one first which is in 3 parts: Part 1 being more of the history and math stat side of things, number 2 discussing how these methods translated and changed from the computer age, and number 3 offers amazing tips on how to code these things in R and goes over many methods developed around the turn of the century (which you'd notice Statistical Inference was published before the turn of the century). Things like neural nets are touched on, bootstrap, great coding stuff. I'm an undergrad rn going into math stats next semester and took into to math stat last semester and the coding aspect is real important for data scientists and something I want to get gud at before going into graduate school; even though the theory side of things typically is much harder to understand. I got this rec from- WAIT LMAOOOOOOOOOO THIS CHANNEL RECCOMMENDED THAT BOOK TO ME, TOO askjdahkjsdakjsdaksjhaskjdk. If you want a more in-depth and accurate review, I recommend the video "Learn Statistics from the Best Professor" on this channel. The introduction and first chapter so far has really great connections with inference!
@@xvzf8147 Statistical Learning by Daniela Witten, Gareth M. James, Trevor Hastie, Robert Tibshirani? Could you name some books theoretical or practical that you consider key to learn data science / statistics properly?
There are some printing mistakes in Chapter 10. E.g. page 489, Theorem 10.3.1 proof has wrong expression for -2*log(lambda(x)). Some of the proofs in the chapter are not complete, they are sketchy, with just a mention of Slutsky's theorem.
Is this really graduate level? The material looks like the material from my undergrad probability textbook, however, it just looks as if the authors wanted to explain it in a more "mathematically-complete" way (I made up that term bc I do not know if there is a better term for it. I think ppl like to say "rigorous," however, I have found that the word "rigorous" is losing its meaning)
Yes, rigorous does not mean harder. It means the logic is tighter and foundations build from axioms. It is the difference between the limit/epsilon approach to calculus (rigorous) vs d(x^n) = nx^(n-1) (algorithm). This books starts with the Axioms of Probability and builds from there.
I have read about that text, but I never read the text. I have many more mathematical textbooks that would compliment Casella better than Wackily. With this topic, I found that multiple angles is helpful so I would not try to get the 'right' book, get a few good books.
This is what I imagine a math stats book would like like if Euclid wrote it. Definitions/Terminology for pages then content. If that is how you like to learn then go for it.
Yh I don’t know man… i could really need this book. I’m learning for my statistics exam and want to smash my pc against the wall. I could instead buy the book and throw the book against the wall every time I’m about to freak out
There too much good book in statistic, i have 25 years from last static reading, but i still like it, i specialized on other branch of mathematics(complex analysis and differential equations)
This is why I love the internet. To put this in context, if you search you'll see that this is one of the textbooks used for the "Statistical Theory" course at Oxford University because it's listed in the lecture notes online. So it's rigorous enough for arguably the top maths university in the UK but it's "not nearly rigorous enough" for some anonymous internet commenter for reasons totally unexplained, also without recommending anything else instead.
