In hope I am not misleading you, I guess it is the STANDARD t-distribution which requires a STANDARD normal R.V. populating the numerator (*). What I do not fully get is how beta* , defined as "a value we got from our LS estimators", serves as a statistic. I think it should be beta_LS_{hat}, much like denominator which contains a statistic. But then again, I do not completely follow the lecturer's intention using this notation. (*) www.statlect.com/probability-distributions/student-t-distribution
Question orthogonal to the video content: may I ask what sort of device you use for the screencasts? I plan to do some on-line tutoring and I have a Wacom tablet, but I find it quite difficult to write legibly with the Wacom pencil. It just isn't the same as paper. May I ask what you use? (I've already told you this, by the way, but the videos are fabulous).
Hey someone could explain to me why we don't use the estimation of var(Beta ls) to approximate the distribution of Beta hat instead of creating a t distribution ?
beta star is the application of beta hat on a particular sample . For one beta hat, you should have several beta star: one by sample . beta star is a number , beta hat is the ols estimator .
Sometimes I wonder why I am still going to school when I can just watch Ben Lambert's video.
I think the numerator has to be beta* - beta_true, because the t distribution requires a standard normal distribution in it's numerator.
In hope I am not misleading you, I guess it is the STANDARD t-distribution which requires a STANDARD normal R.V. populating the numerator (*).
What I do not fully get is how beta* , defined as "a value we got from our LS estimators", serves as a statistic. I think it should be beta_LS_{hat}, much like denominator which contains a statistic. But then again, I do not completely follow the lecturer's intention using this notation.
(*) www.statlect.com/probability-distributions/student-t-distribution
Paying 9k year for University and here I am
the videos are amazing ! This is Quality content ! Great Work !
if N is great enough, why don t we use a normal N(beta_true, sigma_tilda_square) for inference ?
What is sigma_tilda_square?
Question orthogonal to the video content: may I ask what sort of device you use for the screencasts? I plan to do some on-line tutoring and I have a Wacom tablet, but I find it quite difficult to write legibly with the Wacom pencil. It just isn't the same as paper. May I ask what you use? (I've already told you this, by the way, but the videos are fabulous).
Hey someone could explain to me why we don't use the estimation of var(Beta ls) to approximate the distribution of Beta hat instead of creating a t distribution ?
whats the difference between beta_star and beta_hat?
beta star is the application of beta hat on a particular sample . For one beta hat, you should have several beta star: one by sample . beta star is a number , beta hat is the ols estimator .
I was learning string theory and here I am (i.e. I ALWAYS end up in this channel lol)
Use simple notations and expressions
?