Ontology: is "input data" the same as "exogenous data"?

[stap-m]

Description of the issue

This question was raised in #483.
If "exogenous data" and "input data" are not synonyms, we'd need a definition for input data. Also, we should decide which is the class that is needed to finish #483.

Definition of exogenous data: _Exogeneous data is a data item whose quantity value is determined outside of a model and is imposed on a model._

Ideas of solution

If you already have ideas for the solution describe them here

Workflow checklist

• [ ] I am aware of the workflow for this repository

[Vera-IER]

In my understanding input data and exogenous data are not the same. So exogenous data is _a data item whose quantity value is determined outside of a model_, which means that is calculated outside the model, maybe with another model. I would not consider calibration data or statistical data as exogeneous data (for example the installed capacity of solar power in the base year). I would say that exogenous data is a part of all input data.

[akleinau]

@Vera-IER your explanation sounds rather like input data is a part of all exogenous data?

[Vera-IER]

No input data is in my understanding all data which is fed into the model, this could be statistical data and exogeneous data.

[akleinau]

hmm, when we put exogeneous data under input data I'm thinking about putting endogeneous data under output data? Or are those two the same?

[stap-m]

Regarding the definition of exogenous data, I don't really see a difference to what @Vera-IER calls _input data_...

[han-f]

hmm, when we put exogeneous data under input data I'm thinking about putting endogeneous data under output data? Or are those two the same?

There could be endogenous data that is not explicitly spit out by a model, no? I.e. the output a model produces might only be a subset of what is available and endogenously determined ...

[han-f]

In my understanding input data and exogenous data are not the same. So exogenous data is _a data item whose quantity value is determined outside of a model_, which means that is calculated outside the model, maybe with another model. I would not consider calibration data or statistical data as exogeneous data (for example the installed capacity of solar power in the base year). I would say that exogenous data is a part of all input data.

I agree with @stap-m, and think that any data that is fed into a model from the outside - be it calculated data or "empiric" facts/data would be exogenous to the model and thus exogenous data / input data.

But writing this I come to think: there is plenty of exogenous data in the world of course and only a subset of this is actually fed into a model :-)

But for our purpose my feeling would be they could be synonyms as we are only interested in that specific data that enters the model from the outside, be it calculated by another model or simply a statistical value, fact or qualitative fact translated into an input the model can take.

[sfluegel05]

From an "inside-the-model" point of view it makes sense to see exogenous data as a subclass of input data, because only the exogenous data which is fed into the model is actually relevant. From an "outside-the-model" point of view input data is a subclass of exogenous data, because there you see a the data which is not fed into the model.

So the question is: Is the distinction between input and exogenous data relevant? If not, we could use them as synonyms. If it is relevant, it would probably be best to have no direct relation between the classes. Exogenous data could be "a data item whose quantity value is determined outside of a model" and input data could be "a data item whose quantity value is imposed on a model" (basically splitting the current definition of exogenous data in two)

[l-emele]

So the question is: Is the distinction between input and exogenous data relevant? If not, we could use them as synonyms.

But for our purpose my feeling would be they could be synonyms as we are only interested in that specific data that enters the model from the outside, be it calculated by another model or simply a statistical value, fact or qualitative fact translated into an input the model can take.

I agree here to @sfluegel05 and @han-f. Data that exists but is not input into the modelling exercise is irrelevant here. So I vote for exogeneous data and input data as synonyms.

here could be endogenous data that is not explicitly spit out by a model, no? I.e. the output a model produces might only be a subset of what is available and endogenously determined ...

I agree here again with @han-f : Models typically do not spit out all data of their internal calculations. So I would argue that endogenous data is the data of the model that is in principle accessible while the output data is the subset of the endogenous data that is actually spit out by the model.

[stap-m]

We have: _Endogeneous data is a data item whose quantity value is determined by a model._

Thus we coud define _Output data is endogeneous data that is the result of a model calulation._ ? Does this conform to your idea @han-f @l-emele ?

[l-emele]

I agree to your proposed definition.

[han-f]

Thus we coud define _Output data is endogeneous data that is the result of a model calulation._ ? Does this conform to your idea @han-f @l-emele ?

I think this reads good, but I am not sure whether it also communicates that the data is "spit out", i.e. presented to the "outside world". Could this be an idea?:
_Output data is endogeneous data that is the determined by a model calculation and presented as a result._

[l-emele]

_Output data is endogeneous data that is the determined by a model calculation and presented as a result._

I can live with both definition proposals, but I prefer this one.

[l-emele]

Btw, the dictionary says it is _endogenous_ and not _endogeneous_ (without the last e).

[stap-m]

Ok, nice. Then I'll implement.

[han-f]

This is weird, the notifications for this issue went straight to my spam folder, while others come through.