5
$\begingroup$

I stumbled on to these following two slides (slides 21 & 22 on a machine learning tutorial found here):

Scatterplot.. enter image description here

The first is obviously an $x,y$ scatterplot of height and weight.

But what is the second? Looks like two component parts representing the original data, but I can really tease out what is going on. Any insight (or reproducible R code, or even Pseudo-code) on what the heat map is representing, and what the white lines are interpreted as would be great.

$\endgroup$
4
  • 4
    $\begingroup$ A guess: The heatmap corresponds to the ellipsoidal contours of a multivariate Gaussian density obtained by fitting the data (probably by maximum likelihood or similar). The white contour lines correspond to the conditional (Gaussian) density obtained from conditioning on the given height, so, e.g., the bottom line is the density for weights given a height of 140 (cm, I'd guess). Do you have access to the underlying data? $\endgroup$
    – cardinal
    Commented Sep 18, 2013 at 16:29
  • $\begingroup$ Unfortunately, I only have the scatterplot visually.. I think I can see the conditional distributions in the white lines and the max likelihood seems a good guess at the heat map. thanks $\endgroup$
    – T. Scharf
    Commented Sep 18, 2013 at 16:44
  • 3
    $\begingroup$ If the slides were done by the person who is named at the bottom of the slides, why not ask him what they are? It's likely to be more accurate than the out-of-context guesses - however plausible they may be - that we could make. $\endgroup$
    – Glen_b
    Commented Sep 19, 2013 at 5:18
  • $\begingroup$ The white horizontal lines are identical - just shifted to the right as a function of vertical position. I suspect they traverse the semi-major axis of the principal ellipsoid. $\endgroup$ Commented Apr 7, 2014 at 22:52

1 Answer 1

3
$\begingroup$

A guess: The heatmap corresponds to the ellipsoidal contours of a multivariate Gaussian density obtained by fitting the data (probably by maximum likelihood or similar). The white contour lines correspond to the conditional (Gaussian) density obtained from conditioning on the given height, so, e.g., the bottom line is the density for weights given a height of 140 (cm, I'd guess).

-cardinal

If the slides were done by the person who is named at the bottom of the slides, why not ask him what they are? It's likely to be more accurate than the out-of-context guesses - however plausible they may be - that we could make.

-Glen_b

$\endgroup$
1
  • $\begingroup$ I've copied these comments as a community wiki answer because they are, more or less, answers to this question. We have a dramatic gap between answers and questions. At least part of the problem is that some questions are answered in comments: if comments which answered the question were answers instead, we would have fewer unanswered questions. $\endgroup$
    – mkt
    Commented Jun 21, 2019 at 18:38

Your Answer

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

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