On first glance, seems to be moderately advanced, somewhat unusual, at least fairly competent, maybe better than that.
Starts with a typo for expression (1); calls it (5) where it is only on page 5.
Starts with a classification problem and seems to get some interesting and likely useful mathematical (probabilistic, statistical) results for what I've seen elsewhere only with heuristics.
Does prove the Lindeberg-Feller version of the central limit theorem.
I would have guessed that there would be some material on exchangability and/or resampling but didn't see anything.
Looks at the Breiman LASSO technique in regression.
He seems to do a lot with one of my favorite authors and papers:
Talagrand, M. (1996). A new look at independence. Annals of Probability 24, 1-34.
Bro watchu mean moderately advanced? Outside of your PhD dissertation where have you seen more rigorous treatments? I'm no stats schlep, but holy shit this is no walk in the park. Also, am glad you commented, only reason why I actually took a close look at it and it is quite good!
Interesting. The lecture notes don’t look like point estimation and there’s not much hypothesis testing. Course seems much more probability than statistics.
Starts with a typo for expression (1); calls it (5) where it is only on page 5.
Starts with a classification problem and seems to get some interesting and likely useful mathematical (probabilistic, statistical) results for what I've seen elsewhere only with heuristics.
Does prove the Lindeberg-Feller version of the central limit theorem.
I would have guessed that there would be some material on exchangability and/or resampling but didn't see anything.
Looks at the Breiman LASSO technique in regression.
He seems to do a lot with one of my favorite authors and papers:
Talagrand, M. (1996). A new look at independence. Annals of Probability 24, 1-34.