@ 17:52 or so... I did not learn that the SBC is the same as BIC. I have BIC as Sawa Bayesian IC and SBC as Schwarz Bayesian IC...however this is an excellent lecture. The presenter makes me wish I was again in those stuffed 60 seat under grad classes. (Too immature to appreciate those years ago...but fortunate that such ideas are things I find and maintain a healthy curiosity about.)
should not the null hypothesis at 3:15 be H_0: (beta_k | all other betas in the model) = 0 ? since if we remove the other factors from the model, the t-stat will change.
Hello Chris, thanks so much for your informative video! Can Mallows Cp be a part of an iterative forward stepwise feature model selection, by selecting the features with p-value, R^2 and Cp combined on each iteration?? Thanks so much.
@ 17:52 or so... I did not learn that the SBC is the same as BIC. I have BIC as Sawa Bayesian IC and SBC as Schwarz Bayesian IC...however this is an excellent lecture. The presenter makes me wish I was again in those stuffed 60 seat under grad classes. (Too immature to appreciate those years ago...but fortunate that such ideas are things I find and maintain a healthy curiosity about.)
should not the null hypothesis at 3:15 be
H_0: (beta_k | all other betas in the model) = 0 ?
since if we remove the other factors from the model, the t-stat will change.
Hello Chris, thanks so much for your informative video! Can Mallows Cp be a part of an iterative forward stepwise feature model selection, by selecting the features with p-value, R^2 and Cp combined on each iteration?? Thanks so much.
So the CP in a full model is p (number of parameters)? Sorry but I don't speak english
Yes. p is the number of parameters.