Done. Actually, I just dropped the row-vector bit. Doesn't seem necessary and
the "column vectors stored as matrix rows" issue is (a) well-known and (b)
generic to multivariate methods.
> On 14 Nov 2014, at 10:09 , peter dalgaard wrote:
>
> Right. (Well, I suppose that you could transpose x a
Right. (Well, I suppose that you could transpose x and Lambda, but then it
would have to be x = f Lambda + epsilon and all the dimensions would be wrong.)
I'll fix this in R-devel.
-pd
> On 14 Nov 2014, at 00:13 , John Maindonald wrote:
>
> <<<
> Thus factor analysis is in essence a model fo
<<<
Thus factor analysis is in essence a model for the correlation matrix of x,
Σ = Λ'Λ + Ψ
>>>
This should surely be Σ = ΛΛ' + Ψ
Also line 3 under “Details” says
<<<
for a p–element row-vector x, …
>>>
x is here surely a column vector, albeit the transpose of a row vector
from the data matrix.
