That is to say: a graph with multiple feature variables showing different ellipses plots for class groups. What do the ellipses actually tell us? How are the ellipses graphed? What does overlap among ellipses mean? What basic graphing theory techniques are ellipse graphs based on?
I look at ellipse graphs and just think: huh – it looks like clustering but I know it isn’t clustering. I have no idea what this is saying and it is awful confusing.
[Example](https://machinelearningmastery.com/machine-learning-in-r-step-by-step/)
Scroll to section 4.2. The first plot there is the one I am asking about.
In: Mathematics
Like [this?](https://lh3.googleusercontent.com/proxy/FAdDfPz3j3dUoc5GrtCo0G-lVekqn4JqAlbx7CaDO5S5HlJLdUKp4QG3eDoW7Y6tTrDDimHAkiU34_-ATSZZUMlddUsFNgVTn73IElHTfk0vRmucU2lzz2PJ2IxxQos_fb7a91Dm-iAgNOmZBWaN81eq-Wo) That type of graph is less for quantitative consumption, and more for qualitative consumption. Here you’re seeing that rubbers/composites/metals are tougher than porcelain/glass/ceramic and that ceramic/metals/brick are stronger than wood/foam/rubber as classes. If you wanted specific data about a particular type or wood or metal, you need a different chart. This one just shows the ranges.
This graph is elliptical on two axes. You can find ones that are only in one axis as well.
Latest Answers