Honestly, I think PT:TLoS is probably best for those who already understand Bayesian statistics to a fair degree (and remember their calculus). I’m currently inching my way through Sivia’s 2006 Data Analysis: A Bayesian Tutorial and hoping I’ll do better with that than Jaynes.
I think PT:TLoS is probably best for those who understand frequentist statistics to a fair degree. He spends a whole load of the book arguing against them, so it helps to know what he’s talking about (contrary to his recommendation that knowing no frequentist statistics will help). The Bayesian stuff he builds from the ground up, calculus is all that’s needed to follow it.
Jaynes begins it with a caution that this is an upper undergrad to graduate level text, not knowing a great deal of probability in the first place, I stopped reading and picked up a more elementary text. What do you think are the core pre-reqs to reading Jaynes?
I’d agree, with the exception that chapters one and five (and maybe other sections) are great for just about anybody to get a qualitative understanding of Jaynes-style bayesian epistemology.
Honestly, I think PT:TLoS is probably best for those who already understand Bayesian statistics to a fair degree (and remember their calculus). I’m currently inching my way through Sivia’s 2006 Data Analysis: A Bayesian Tutorial and hoping I’ll do better with that than Jaynes.
I think PT:TLoS is probably best for those who understand frequentist statistics to a fair degree. He spends a whole load of the book arguing against them, so it helps to know what he’s talking about (contrary to his recommendation that knowing no frequentist statistics will help). The Bayesian stuff he builds from the ground up, calculus is all that’s needed to follow it.
Jaynes begins it with a caution that this is an upper undergrad to graduate level text, not knowing a great deal of probability in the first place, I stopped reading and picked up a more elementary text. What do you think are the core pre-reqs to reading Jaynes?
I have no idea—I’ll tell you when I manage to satisfy them!
I’d agree, with the exception that chapters one and five (and maybe other sections) are great for just about anybody to get a qualitative understanding of Jaynes-style bayesian epistemology.
Ah, yeah—chapter 5 is pretty good. (I recently inserted a long quote from it into my Death Note essay.)