Good.

Okay, as usual, let's start with a small quiz about last week.

So last week in the beginning we talked a little bit, or my colleague talked a little bit about Gaussian classifiers,

and this is where also the first question comes in.

So here the question is, you have, okay, one number is wrong here, the first one is of course this one,

and the second one is that one.

So you have here three different scenarios of a Gaussian classifier

where different Gaussians are used to model these classes.

So either distinct diagonal covariance matrices,

identical covariance matrices, or distant full covariance matrix.

So our quest is now to assign which one is which.

Can I make that full screen?

Ah, yeah, at least partially.

Okay, what do you think?

Maybe some more want to vote here.

Okay, not so many today here.

So let's see what you got.

So, okay, one of three got it right or something like that.

Okay, let's go through it a little bit.

So we had that, so this one is easy, this one must be,

let's go to the options.

So these are where we have totally identical covariance matrices.

You see also that here the shapes of the covariance matrices,

of these ellipses, so these covariance matrices,

you can always draw them as ellipses, and you see them in two dimensional.

So you see that these are totally equal, so this must be here with the identical one.

So now the question is which one here of these two here is the distinct full covariance matrix

or a distinct diagonal covariance matrix.

And diagonal means that the covariance matrix shape can only expand in the dimensions of the axis.

So the second one would be the diagonal covariance matrix.

And the last one here is really, then the shapes can also vary,

and you see it's not along an axis, so this must be here the full covariance matrix here.

Okay, let's go to the next question.

The next question here is about K-nearest neighbor classifier,

I just made three small statements and you have to choose the one which is correct.

So there's only one single one correct here.

I can also put that here again.

No, it doesn't work anymore.

But it's QYN3 for vote.

Oh, excuse me.

Thanks for letting me know.

Okay.

Maybe that is the reason why that doesn't work.

No.

Don't ask me why that is the case.

Sometimes it works, sometimes not this QR code.

Does someone need the QR code again?

Yeah, you would need it.

Okay, let me see how I get it.

Here is the link.