You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Partially done here 2336af0, but by creating a new plotDiff generic for "by" smooths. For 'fs' smooths it might actually make sense creating a layer instead.
This might be a quite useful layer, see
http://www.fromthebottomoftheheap.net/2017/10/10/difference-splines-i/
To calculate the standard error for the differences in smooth we need:
cov(X1b1, X2b2) = X1 %*% Sig11 %*% t(X1) + X2 %*% Sig22 %*% t(X2) - X1 %*% Sig12 %*% t(X2) - X2 %*% Sig12 %*% t(X1)
where Sig11, Sig22 and Sig12 = Sig21 are blocks of cov(b1, b2).
The text was updated successfully, but these errors were encountered: