Terrific! I'd seen the theory, but it helps to see it used with real data. I'm getting my MS in Applied Statistics from UDC, and my thesis will be on Bayesian prediction intervals on Poisson processes. This helps a lot. Thanks so much!
In the binomial model, there are really "n" data points, so this does have multiple data points. Perhaps you want Bayesian parameter estimation in a normal model? th-cam.com/video/L5pg_cx2nNU/w-d-xo.html
Thank you very much! May I ask why is the Beta distribution used as the prior - is it just simply for the ease since the posterior will have beta distribution after multiplying the likelihood function (binomial) with the beta?
Yes, precisely because the posterior will have a beta distribution and thus this prior is "conjugate". You could use a different distribution, but then you would have more mathematical work to do to find the posterior.
Terrific! I'd seen the theory, but it helps to see it used with real data. I'm getting my MS in Applied Statistics from UDC, and my thesis will be on Bayesian prediction intervals on Poisson processes. This helps a lot. Thanks so much!
Please keep posting more bayesian statistics (probably advanced bayesian statistics) videos. Very much helpful for my studies at unsw sydney!
I might have the opportunity to this fall as I am teaching an advanced Bayesian stats class.
@@jaradniemi thanks!
I don't know if you still make videos but I just want to mention that this video was extremely helpful, bless you man and hope you are doing well.
I just started making more videos, although perhaps at a more basic level than you need.
This is an awesome lecture. Thank you
Excellent explanation, thanks Jarad
How would the 95% HDI is between (57, 74) according do the distribution plot?
Thanks for creating this content. Really helpful! Would be better if you show Bayesian estimation works with multiple data points :)
In the binomial model, there are really "n" data points, so this does have multiple data points. Perhaps you want Bayesian parameter estimation in a normal model? th-cam.com/video/L5pg_cx2nNU/w-d-xo.html
@@jaradniemi Thanks for the reply! I meant the examples for continuous distributions. Exactly what you just shared.
Thank you very much! May I ask why is the Beta distribution used as the prior - is it just simply for the ease since the posterior will have beta distribution after multiplying the likelihood function (binomial) with the beta?
Yes, precisely because the posterior will have a beta distribution and thus this prior is "conjugate". You could use a different distribution, but then you would have more mathematical work to do to find the posterior.
Can i have the copy of your presentation?
Clearly explained. Thanks Jarad!
Wow, well explained. Sound like you are a professor
Very Good mini-lecture! Thanks a lot!
You answered a lot of my questions, thanks
Very helpful lecture, thanks.
Thanks a lot. Very helpful.
Can you organize the videos properly with numbers? THe videos are great but hardly organized!
They're organized here th-cam.com/play/PLFHD4aOUZFp0Xhzd5j1nWnExD54xJfnJX.html
Please I need your email to contact with you My PHD in blind estimation and I need ask you on it
great work!
Great lecture, thanks
Thankssssssssssssssss