how can I find the distribution of Yi = β1Xi + εi, when we know Xi is normally distributed and mean and standard deviation are unknown and εi is normally distributed with (0,1)? Can we assume it is also normally distrubuted?
Asked
Active
Viewed 28 times
0
-
Linear combination of normally distributed random variables I think is also normally distributed. – Fiodor1234 May 09 '21 at 20:26
-
1@Fiodor Only when you also assume all the variables have a multivariate normal distribution. See https://stats.stackexchange.com/questions/30159. – whuber May 09 '21 at 20:53
-
1@whuber Really englihting post! So, as I understand it $Y_{i}$ could be assumed normal if we it already holds that $(b_{1}X_{i},\epsilon_{i})$ is multivariate normal, based on the second anwser of the post? – Fiodor1234 May 10 '21 at 09:24
-
@Fiodor1234 Yes, that's right. And since we usually assume all $\varepsilon_i$ are independent of all $X_j,$ that implies the entire collection of $X_j$ and $\varepsilon_i$ have a multivariate Normal distribution when the $X_j$ are multivariate Normal and the $\varepsilon_i$ are multivariate Normal. – whuber May 10 '21 at 14:32