Timeline for Econometric Model for Vector-Valued Variables
Current License: CC BY-SA 4.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 27 at 22:07 | comment | added | mark leeds | Darek: I was just pointing out that it looks similar to multivariate regression. minginator seems to have a good grip on this material ( I don't ) so it's best to let him help you. And thanks to minginator for expressing my comment more clearly. | |
| Nov 26 at 16:47 | comment | added | minginator | Darek, you are referring to now is multiple regression, @markleeds is referring to an equation of the form $Y = XB + \varepsilon$ where $Y$ and $\varepsilon$ are $n \times p$ matrices, $X$ is $n \times q$ matrix and $B$ is $q \times p$ (I use $n$ to denote sample size, if you take $n = 1$ and the transpose you would get the model in your post for a single observation). You may find more if you search for Seemingly Unrelated Regressions (SUR), or e.g. here (bashtage.github.io/linearmodels/system/…). | |
| Nov 26 at 16:03 | history | edited | Darek | CC BY-SA 4.0 |
added 16 characters in body
|
| Nov 26 at 16:01 | comment | added | Darek | @markleeds In multivariate regression on left hand side we have $Y$ which is a $n \times 1$ vector. In this model $Y$ is form as $n$ vectors each one $k$ dimensional. Thanks - you are right about $\epsilon$ | |
| Nov 26 at 15:58 | history | edited | Darek | CC BY-SA 4.0 |
added 16 characters in body
|
| Nov 26 at 15:36 | comment | added | mark leeds | I haven't seen that before myself but it's pretty similar to multivariate regression. See this for the similarities. brilliant.org/wiki/multivariate-regression Note that you can't have a scalar $\epsilon$ if you have a vector response. | |
| Nov 26 at 14:12 | review | Close votes | |||
| Nov 27 at 10:38 | |||||
| Nov 26 at 13:56 | history | asked | Darek | CC BY-SA 4.0 |