10 - Knowledge Discovery in Databases [ID:53466]
50 von 720 angezeigt

Okay, before we start with our real lecture, we had one question after last lecture, but

yeah, it's relevant for everybody. So I will answer it today because the exam is only one month away.

I think it's important to talk about which helpers are allowed in the exam. Just like last semester,

we will allow a one-sided thin A4 paper as a cheat sheet during the exam. It's very important that you only have handwritten notes in that,

so no printing allowed and also not to write on an iPad or a Surface or whatever and print it out because we want to have some repetition while creating these cheat sheets by the students.

So while creating the cheat sheet, you will repeat the most important things because you have to select things to put in the cheat sheet and things like that.

And this repetition is also a very good learning experience in our opinion. So you will have to do it by hand and on your own.

And we will collect those cheat sheets after the exam, so you can't hand them over to a second student, so this student also has this learning experience.

We will not allow calculators, however, in the exam. I think there was a question in the forum regarding that.

Of course, if you take a look at the current, for example, classification exercises, there have been pretty complicated computations in that.

We do not expect you to be able to compute a logarithm of one-fifth in your head, of course. That's nothing we will expect of you.

There will be computations in the exam, however, they will be of a pretty low level, so everybody should be able to do them in their head if they can do basic math.

But there will not be any difficult computations during the exam.

That's also a pretty good hint on which formulas might be relevant for the exam and which won't be relevant.

Because if we, for example, take a look at a Gaussian distribution, a pretty complicated formula, even with the simplest values, nobody will be able to calculate a Gaussian distribution in their head.

That's something that we are aware of. So there's nearly no probability, a probability close to zero, that we will ask you for a Gaussian distribution in the exam, because it's hard to compute in your head.

Also, if we have, for example, a decision to the induction, we have to make sure you can calculate those things in your head.

So by not having a calculator might be worrying for you.

It's actually something very good for you because it lowers the amount of things we can ask for in the exam.

You just have to think about that this way.

Besides that, in the exam, we will also ask you to bring a correction roller because most of our exams, our exam questions will be multiple choice or single choice questions.

We will mark them in the exam, whether they are single choice or multiple choice.

And we will scan your exams and automatically correct those questions, single choice and multiple choice.

That's done via an object recognition algorithm, which requires you to mark a certain spot on the exam.

So there is a chance that you will have these little rectangles on the exam for yes and no.

Or of course, there might be more options than two options, but let's go with yes and no.

And there are little rectangles besides that.

You will have to fill them out fully for them to be counted as marked.

However, there might be a case where you mark something and then realize it's not the correct answer.

Then of course you have to cross it out.

If you would cross it out this way, our object recognition algorithm wouldn't recognize that.

So you would need something to wipe the space out again because then it's recognized as not marked.

And that can be done via a correction roller, not correction fluid in the exam.

But those are things I will talk about in the last lecture of the semester again.

But if you have to order a correction roller, you might think about that in the next time.

Maybe I put a little note on the forum as well and things like that.

So allow helpers on one side of the four cheat sheet handwritten, not printed, will be collected after the exam.

No calculators.

And we recommend a correction roller to correct wrong answers, wrongly marked answers.

Those are the most important things to talk about the exam today.

Of course, in the last lecture, we will talk about the exam in some more detail afterwards.

And if you have questions, of course, you can always ask them in the forum or things like that.

OK, are there any questions regarding the exam right now?

OK, I don't guess so.

Then we can go back to clustering.

We stopped in the chapter in the subsection of hierarchical clustering,

one very specific sub kind of clustering methods where we can go either from small clusters to bigger clusters,

combining them together through a clomerative clustering or can split them via divisive clustering.

We stopped at Birch.

Birch is a method called balance, iterative, reducing and clustering using hierarchies.

Teil einer Videoserie :

Zugänglich über

Offener Zugang

Dauer

01:33:13 Min

Aufnahmedatum

2024-07-01

Hochgeladen am

2024-07-03 09:16:03

Sprache

en-US

Einbetten
Wordpress FAU Plugin
iFrame
Teilen