Hi @Shihab_Khan
Are you referring to the discussion on this topic? This is a work-around that should in principle produce the right results, but is not officially supported by UQLab.
However, I think that it should also work in your case, and to me it looks like everything went well inside the analysis. Do you get unexpected results?
The issues you mention can be easily explained:
The number of model outputs are 2, although I agree that
1
to
1
2
to
2
should be better formatted as
1
to
2
The number of observations in the UQLab sense is only 1, because Data(1).y contains only one row. I suggest you have a look inside uq_inversion_likelihood, to see how UQLab compares your data to the model outputs.
Let me know if you have further questions.