4 ms·
Good point, that really is the key. However, itjust turns it into the almost equivalent problem of: why assume Gaussian error terms, and the rest of the articl
by masterjack 11y ago
Good point, that really is the key. However, itjust turns it into the almost equivalent problem of: why assume Gaussian error terms, and the rest of the article does a good job at motivating this.
- yummyfajitas 11y agoGaussian is sometimes a good approximation if the error is a sum of independent thin-tailed variables (due to the c.l.t.). That is all. The article does not, in my opinion, motivate this very well.