>>11663980I'll attempt one right now, but it will be more application focused rather than hardcore theory... My assumptions for prior knowledge:
-Feller Volume 1
-Some sort of measure theory course
Then the topics you can look at as a beginner are(chronological):
Basics-Statistics by Pisani and Friedman. This is like the baby version that cover very beginner concepts. However, you need to understand them properly to not end up like a CS-oid with retarded modelling skills.Skip the probability parts.
Beginner-Statistical Models by Friedman. This is where you start doing some modelling, and should be within your reach after the 1st book. You can look at Lasso and ridge in more detail in the original articles.
Intermediate-Design and Analysis of Experiments by Montgomery. There is overlap, but this book really gets you all you need to know for designing experiments. You need to know ANOVA's very well, and this is dedicated to that
Bayesian/Frequentist split-At this point,if you are a Bayesian, you might want to get into Bayesian stats. Here without any question I recommend Gelman, Bayesian Data Analysis. Somewhat hard, but allows you to access Bayesian methods very well later on.
Niche topics:
Extreme Value Theory-This is really important if you aim to do some sort of insurance career, but here you should look at Modelling Extremal Events by Embrechts, this is usually accepted as a classic reference.
ML-Others have mentioned into to statistical learning by tibshirani, this is good. For modern methods, refer to goodfellow, not sure they came out with anything better.
Time Series-This topic is honestly such a clusterfuck in terms of rigour and presentation that there really isnt anything that good here. My guess is Brockwell and Davis but this field is composed of complete voodoo techniques so not that important.
Power Law stuff- Goes nicely with Embrechts, pick Taleb Statistical Consequences of Fat Tails
