Lecture75 (Data2Decision) Bayesian Regression, part 2

[@PedroRibeiro-zs5go]

Thanks!! Very nice explanation of the topic!!

[@marcpalaci9224]

nice lecture. I have one small disagreement with your opinion about just using the frequentist approach if you have no knowledge of a prior. As much as bayesian regression is great due to the incorporation of prior knowledge... an arguably better feature is the embedded occam's razor effect (very strong resilience to overfitting). Hence for that reason using non informative priors may still be very beneficial as opposed to using pure MLE.

[@amrendrasingh7140]

Very concise..

[@test-mm7bv]

actual examples with calculations would making conversations about interpretation more meaningful and useful. the 2 videos on bayesian regressions so far sound like a humanities class. hope the next lecture goes into mcmc, map, or other bayesian parameter estimation approaches so i can actually use this to solve problems (rather than blovate about philosophy).