You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Plugin for evaluation of response from multiple input parameters
Motivation
The existing parameter-response correlation plugin allows for correlation of only a single parameter with a response (the ParameterResponseCorrelation plugin).
It is quite common that just a small number of parameters are the key drivers for a specific response, e.g. the total production over the lifetime of a well. A limited number of key parameters is much easier to handle, both when making a decision, and to sort out which uncertainties that are worth putting more work into.
Some suggestions:
Evaluation / literature search of applicable methods to use.
Consider linear and non-linear methods.
Test the most promising ones on the available test data set.
Implementation and delivery:
Implement chosen method(s), either using an existing (well maintained) open source library, or from scratch.
User friendly visualization plugin. Should at least feature: selection of response and an "easy to understand" view of key parameter combinations for selected response together with there "score" (quality of fit)
Keep in mind during implementation: How computationally expensive is the chosen method? Can we calculate on request, or do we have to pre-calculate?
Additional ideas you might have that can provide additional insight to parameter - response correlations
The text was updated successfully, but these errors were encountered:
Plugin for evaluation of response from multiple input parameters
Motivation
The existing parameter-response correlation plugin allows for correlation of only a single parameter with a response (the ParameterResponseCorrelation plugin).
It is quite common that just a small number of parameters are the key drivers for a specific response, e.g. the total production over the lifetime of a well. A limited number of key parameters is much easier to handle, both when making a decision, and to sort out which uncertainties that are worth putting more work into.
Some suggestions:
The text was updated successfully, but these errors were encountered: