6
$\begingroup$

So, in PCA, we decompose the covariance matrix by its eigenvalues and eigenvectors. I understand that an ellipsoid is fully characterized by the eigenvalues and eigenvectors of a positive definite matrix $A$, and it has an equation $(x-x_c)^T A (x-x_c) =1$, where $x_c \in \mathbb{R}^n$ denotes the center of the ellipsoid. In PCA we are basically trying to fit an ellipsoid to the data from what I understand.

I also understand that $i$-th principal direction of the ellipsoid is given by the direction of the $i-$th eigenvector $v_i$ associated with the eigenvalue $\lambda_i$, sorted out so that $\lambda_1 > \ldots > \lambda_n > 0$. And finally, $\lambda_i$ defines the inverse square of the associated semi-axis, so $r_i = \frac{1}{\sqrt{\lambda_i}}$ is the semi-axis of the $i$-th principal direction, right? Now, wouldn't this mean that $r_1$ is the smallest of all other $r_j$? But that can't make sense I think, because then why would we project our data in that direction over any other if we want to reduce the dimension of our data?

I know there has to be something wrong in my reasoning, I'm in no way suggesting PCA is wrong, but I would appreciate some clarification.

$\endgroup$
3
  • $\begingroup$ If my wording is confusing, I can provide an example. $\endgroup$
    – user45453
    Oct 14, 2016 at 16:23
  • $\begingroup$ Consider the r.v. $\mathbf{x}$ following a Gaussian distribution with mean $\mathbf{\mu} = [0 0]^T$ and covariance matrix $\mathbf{\Sigma} = \begin{bmatrix} 25 & -24\\ -24 & 25\end{bmatrix}$. In this case, the eigenvalues are $\lambda_1 = 49$ and $\lambda_2 = 1$, and the eigenvectors are (for instance), $v_1 = [1,-1]^T$ and $v_2 = [1,1]^T$. Now, clearly if we sample this distribution or if we plot some of its contour lines, we see that the principal direction is indeed given by $v_1$. $\endgroup$
    – user45453
    Oct 14, 2016 at 16:27
  • $\begingroup$ Now, in this example, shouldn't the semi-axis associated with the direction given by $v_1$ be 1/7 that of $v_2$? Because $r_1 = c/\sqrt{\lambda_1} = c/\sqrt{49} =k/7$ and $r_2 = c/\sqrt{\lambda_2} = c/\sqrt{1} = c$, where $c>0$ is just some constant for the contour lines. $\endgroup$
    – user45453
    Oct 14, 2016 at 16:29

1 Answer 1

3
$\begingroup$

I think there are two ellipses that we could consider. First, consider the image of the unit circle with respect to the map $x \mapsto x^T A x$ for PD $A \in \mathbb R^{n \times n}$. It is a standard result that $f(x) = x^T A x$ is maximized over unit vectors $x$ by the unit eigenvector $v_1$ with largest eigenvalue $\lambda_1$. So this means that the ellipse formed by the image of the unit circle with respect to this map has its largest semi-axis as $v_1$ with length $\lambda_1$, and so on for the other eigenpairs. So in this case we clearly have that the eigenvalues give the lengths of the semi-axes and the biggest semi-axis is for the biggest eigenvalue.

But now consider the contour $f(x) = 1$ for any $x \in \mathbb R^n$. Since $A$ is positive-definite we know that $f$ is a paraboloid, so its intersection with a horizontal plane (i.e. the contour) is an ellipse. In this case, we find that the shortest semi-axis is parallel to $v_1$, which makes sense because that's the direction that $f$ grows the fastest so we hit 1 the soonest. The largest semi-axis is $v_n$ since that's the direction in which $f$ is growing the slowest. Plugging $v_1$ in to $f$ we get $f(v_1) = v_1^T A v_1 = \lambda_1 v_1^T v_1 = \lambda_1$, not 1 as required, so the actual vector parallel to $v_1$ is $\frac{1}{\sqrt \lambda_1}v_1$. Does this help?

Bringing this back to PCA, let's say that our data consists of $m$ observations coming iid from $\mathcal N_2(\vec 0, \Sigma)$. Let's draw an ellipse around our data such that every point in the ellipse has likelihood greater than some cutoff $c$. This corresponds to a contour of the likelihood and can be found by $$ c = \frac{1}{2\pi \vert \Sigma \vert}\exp \left( -\frac12 x^T \Sigma^{-1} x\right) \iff x^T \Sigma^{-1} x = -2\log (2\pi c \vert \Sigma \vert) $$ i.e. the ellipse that circles the data is a contour of the quadratic form $g(x) = x^T \Sigma^{-1} x$. Note that $\Sigma^{-1}v = \lambda v \implies \frac1\lambda v = \Sigma v$, so the eigenvectors defining the axes of this likelihood contour ellipse are the same as those for the covariance matrix $\Sigma$ but with inverted eigenvalues.

$\endgroup$
6
  • $\begingroup$ Sorry, I got distracted, but YES, it does help! $\endgroup$
    – user45453
    Oct 14, 2016 at 18:01
  • $\begingroup$ @user45453 Great, I'm glad that was helpful :) $\endgroup$
    – jld
    Oct 14, 2016 at 18:20
  • $\begingroup$ +1. Just to clarify: "This is a cross-section of a paraboloid so it's also an ellipse" - can you expand on this a little bit? $\endgroup$
    – amoeba
    Oct 14, 2016 at 18:51
  • $\begingroup$ @amoeba I expanded that a bit. Does this seem clearer? $\endgroup$
    – jld
    Oct 14, 2016 at 19:11
  • $\begingroup$ Thanks. It's just that I don't know what a "paraboloid" is in $n$ dimensions... but I can see what you mean in 3D so I guess it somehow generalizes to higher dimensions too. $\endgroup$
    – amoeba
    Oct 14, 2016 at 19:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.