Sensitivity analysis should be included in the ontology

[akleinau]

Description of the issue

Sensivities (in the context of sensitivity analysis) should be included in the ontology as they are used in the excel scenario factsheets.

Ideas of solution

Workflow checklist

• I discussed the issue with someone else than me before working on a solution
• I already read the latest version of the workflow for this repository
• The goal of this ontology is clear to me

I am aware that

• every entry in the ontology should have a definition
• classes should arise from concepts rather than from words

[han-f]

I wonder whether we should name the term "sensitivity analysis" instead. I think "sensitivity" is more colloquial (and is also used in other contexts).

My line of thinking:
when someone says: "I ran a sensitivity" , he/she probably implies that he/she modified parameters of a model according to a certain method, ran the model again and created outputs based on the changed parameters. This whole exercise is done to compare the outputs based on the sensitivity settings with the "original" outputs based on the original settings.

The factsheet for PszVIII(scenarios) states for example under "Sensitivities": Analysis for higher population development, economic development, lower energy prices in comparison with lower CO2-price path. -> So maybe the term in the factsheet was a bit fuzzy to start with?
What do the others think?
@stap-m @jannahastings @l-emele @christian-rli @Ludee

[jannahastings]

I wonder whether we should name the term "sensitivity analysis" instead.

This sounds good to me.

[christian-rli]

As someone from not in the field I would not have known what was meant by just "sensitivity". "sensitivity analysis" is more self explanatory and also doesn't seem redundant to me. So it sounds good to me as well.

[akleinau]

sounds like sensitivity analysis is best described as a process? The process of recording outputs of a model based on different sets of parameters that are chosen according to different methods?

[han-f]

sounds like sensitivity analysis is best described as a process? The process of recording outputs of a model based on different sets of parameters that are chosen according to different methods?

I could go with that.

[stap-m]

Process sounds good to me. Create a parent class analysis for sensitivity analysis? I am not sure if this is necessary/helpful, though.

An analysis is more that recording outputs.
The process of comparing different model calculations based on a variation of input parameters (exogenous data items)?
Also include: has part some model calculation?

[akleinau]

I think it would be enough to include an analysis class when we need a second analysis subclass.

Is your def a def for sensitivity analysis or for analysis?

[stap-m]

Ok. It's for sensitivity analysis.

[akleinau]

input parameters is more intuitive for me, so go with that instead of exogeneous data items?

[christian-rli]

input parameters is more intuitive for me, so go with that instead of exogeneous data items?

I feel the same way. Was there a particular reason for naming it exogeneous data items in the first place? It seems very specific.

[akleinau]

we just included that class as well as endogeneous data item in #482 as part of including the classes needed for the oeo factsheets

[stap-m]

Yes, it's exogenous data btw, sorry. Since we already have the class, I'd like to refer to it in brakets.

[akleinau]

I would still leave it out because with the brackets it looks like it‘s just another term for input data and in my understanding those two concepts are not the same?

[stap-m]

We don't have input data included yet... How would you define it?
Exogeneous data is a data item whose quantity value is determined outside of a model and is imposed on a model.

[akleinau]

do we need that class and should create a new issue for input data?

[sfluegel05]

#520 came to the conclusion that input data is the same as exogenous data. So we can use the definition for sensitivity analysis @stap-m suggested:

The process of comparing different model calculations based on a variation of input parameters (exogenous data items)?

[han-f]

#520 came to the conclusion that input data is the same as exogenous data. So we can use the definition for sensitivity analysis @stap-m suggested:

The process of comparing different model calculations based on a variation of input parameters (exogenous data items)?

Reads good to me.

[l-emele]

I agree, too.

[l-emele]

I'll implement.