Mordicus.Core.OperatorCompressors.Regression module

Mordicus.Core.OperatorCompressors.Regression.CompressOperator(collectionProblemData, solutionNames, operatorCompressionInputData)[source]

Computes the offline operator compression stage using the method of POD on the snapshots and a regression on the coefficients

Parameters

  • collectionProblemData (CollectionProblemData) – definition of the training data in a CollectionProblemData object

  • solutionNames (list of str) – names of the solution to be treated

  • operatorCompressionInputData (dict of objects satisfying the scikit-learn regressors API) – input regressor to be fitted

Mordicus.Core.OperatorCompressors.Regression.ComputeOnline(onlineProblemData, solutionName, operatorCompressionOutputData)[source]

Compute the online stage using the method of POD on the snapshots and a regression on the coefficients

The parameters must have been initialized in onlineProblemData

Parameters

  • onlineProblemData (ProblemData) – definition of the testing configuration data in a CollectionProblemData object

  • solutionName (str) – names of the solution to be treated

  • operatorCompressionOutputData ((regressor, scaler, scaler)) – (fitted regressor, fitted scaler on the coefficients, fitted scaler on the parameters)

Returns

onlineCompressedSnapshots; dictionary with time indices as keys and a np.ndarray of size (numberOfModes,) containing the coefficients of the reduced solution

Return type

collections.OrderedDict