I used the Telemac2D software to solve the Shallow Water Equations and got 1000 result files that i want now to post-treat and do some sensitivity analysis with OpenTURNS. The output i mainly want to focus on is the water level. In my case, i can have it's value for different time steps.
My problem is the way of linking the inputs to the output of interest in order to launch a polynomial chaos algorithm, since i don't have an explicit representation of the output by the input (if i had it i could use ot.SymbolicFunction
or PythonFunction
).
In the case where the inputs and outputs are known only from a
Sample
, we may use the following constructor:where
inputTrain
is the inputSample
andoutputTrain
is the outputSample
. The two samples must have the same size (but not necessarily the same dimension). In this case, the library performs several steps before fitting the polynomial chaos:inputTrain
Sample
,This way of creating the chaos is presented in the following page: Fit a distribution from an input sample.