P3:
I had made a kind of circle.
It starts a bit with the reports that are there, of nuisance and so on.
I also thought of making that clear to the subject.
Whether he also thinks that those reports are justified or correct or at least knows about them.
Then... yes, then the model makes some kind of assessment or decision based on the data about that house and those people whether it is possible fraud.
And that if the employee starts looking at it, then the certainty of such a prediction is also there.
That explanation was already there from SHAP.
That that can help a little bit with, yes, how sure are we of this or possibly unsure.
And indeed what is decided for what.
And then the employee can start thinking, okay, so we have to look into this further because it's probably fraud... probably.
And that you also have the data you have about those people checked by the person it is about.
Because maybe it's just a mistake somewhere in the system or it's just not set up properly yet.
That you also submit that kind of hey we now have this about you, is that right?
Without immediately knocking on the door and saying that you are a fraud.
But if it turns out to be something, that you go to the person of the report to present the problem, like this is going on, this is missing, or this is wrong, you have not done this correctly.
And that you will also more or less estimate whether it is fraud or a mistake.
That you might do that more like together with the person in question.
Yes, we have written down... the conversation about the personality, like is it really justified to characterize this as fraud and really impose such a fine whether that fine is justified.
And I had added here not only legally but just in a human sense.
And we've written a bit loosely about people being able to challenge something like that easily when a decision is made.
On the one hand, do the people themselves know that it is possible, that it is allowed, where and how you just want that to be clear.
With minimal administrative hassle.
And even then ultimately the loop for, if you often see that it is actually disproportionate or something, that you actually just adjust the policy or those rules.

R1:
Great.

P4:
Yes, well...
These discussions were let's say we had indeed and that's what I'm trying to do...
Our basis, at least one of the basis, was that we went to see where things could go wrong.
And I actually had two loops the first one was in…
Or we had two loops sorry.
But the decision to check so the first piece is if we are going to check someone at all and the second piece is if we are going to fine someone.
The first piece is very much based on all the data. Where is the house, who is the declarant, is it anonymous.
So all of those things that…
All variables that you should want or be able to do something with.
And... also... the number of reports about a location or the number of reports from a person could be indicative of a risk or something.
So that.
Oh yes, the two loops decision to control and decision to penalty, how often are they correct?
You have to be able to compare them with each other because there is a human side to that decision to fine.
"Knock knock knock' this is not correct and "money" is there really a fine handed out?
Because they will run based mainly on the algorithm to then have less work and grab the top 10 or so.
Well... the... how... very important indeed yes but if it goes wrong then how can you knock on a door because yes that is...
I also just realized that creating things like this, yes, those people are not there for that.
Those people just want to rent out their house and they don't want to say...
The people who come to you, it is their job to do and execute all that.
While you just really only want to understand the basics of the rules.
And then I think you have the difference between that you have done something wrong consciously and that you have done something wrong unconsciously.
Well and somewhere it starts there then it becomes a vague loop and thing.
Because I think you want the person who has insight into the rules and deliberately makes a mistake, you want to fine them proportionately.
And the person who unknowingly makes a mistake or did not understand the rules properly, you may want to fine in a different way.
And then you can have the convenience of intervention, but Joke only wanted a week's vacation.
That was just, yeah...
And then suddenly you are in a paper tiger and you spend a year trying to prove that you live on 1A and not on 1B.
So yes.
Same input.
But yeah.