So let's continue in the text and it's Monday morning so we should discuss briefly the mind

map.

I told you that sometimes I feel bad about always repeating myself with a mind map but

it also shows in the evaluation that students appreciate it, the mind map.

And so let's just do an investment of another 10 minutes to give you the big picture and

I can tell you if you browse through the videos while preparing the lecture and if you just

focus on the mind maps that I do over and over again, then you get a very good feeling

how I think and how I work on these topics because I do not prepare the mind map so I

do this right from my heart, by heart and write things down and that's the way also

I generate questions during the oral exam.

I just think about the topics and then I generate the questions on the fly.

I do not prepare questions for the oral exam.

I just generate them on the fly.

That's important to know.

And I also read in the web, you know, I also of course read all the discussions you do

with how was the oral exam and which questions came up and so on and one guy was writing,

I mean, we can discuss whatever we want.

This guy is asking everything and that's basically telling the truth.

I basically ask everything and I do not focus on certain topics in the lecture, in the oral

exam, in the lecture.

Hopefully I focus.

Good.

Winter, summer semester, winter semester is introducing the core concepts of a simple

classifier with signal acquisition, filtering, computational features, classification and

so on.

In summer semester we have a broader view towards things and in the main focus is basically

the modelling of the a posteriori probability.

So if you think about pattern analysis as we teach it this summer semester, you should

have in mind a cloud and within this cloud there is the P of Y given X where Y is the

class number and X is the feature vector.

Why did I change the notation?

It took me or it was quite tough for me to do this decision but the notation with Y simplifies

many derivations.

Think about support vector machines, think about Rosenblatt's perceptron.

So we have used Y as a variable that gets numbers minus one or one or zero or one and

it's much more compact in writing things down compared to the notation we have used in winter

semester.

So I decided to adapt to the international standard with respect to this notation.

X is a feature vector, Y is the class number.

Good.

Dehydration is a serious problem, especially for elder people so go ahead.

We also applied for a research project where we want to detect dehydration of elderly people

by using special sensors.

The problem is that they dehydrate because they do not drink enough and so you can either

measure it on the skin or you can also measure the activity of the toilet for instance to

check whether people have enough fluid.

Thanks for this pointer.

So we discussed here the a posteriori probability and now of course the question is why I am

so eager to get the a posteriori probability.

Stefan?