1%feature("docstring") OT::LinearLeastSquaresCalibration 2"Linear least squares calibration algorithm. 3 4Available constructors: 5 LinearLeastSquaresCalibration(*model, inputObservations, outputObservations, candidate, methodName*) 6 7 LinearLeastSquaresCalibration(*modelObservations, gradientObservations, outputObservations, candidate, methodName*) 8 9Parameters 10---------- 11model : :class:`~openturns.Function` 12 The parametric function to be calibrated. 13inputObservations : 2-d sequence of float 14 The sample of input observations. 15 Can have dimension 0 to specify no observations. 16outputObservations : 2-d sequence of float 17 The sample of output observations. 18candidate : sequence of float 19 The reference value of the parameter. 20methodName : str 21 The name of the least-squares method to use for the calibration. 22 By default, equal to *QR*. Possible values are *SVD*, *QR*, *Cholesky*. 23modelObservations : 2-d sequence of float 24 The sample of output values of the model. 25gradientObservations : 2-d sequence of float 26 The Jacobian matrix of the model with respect to the parameter. 27 28Notes 29----- 30LinearLeastSquaresCalibration is the minimum variance estimator of the parameter of a given model under the assumption that this parameter acts linearly in the model. 31 32The prior distribution of the parameter is a noninformative prior 33emulated using a flat :class:`~openturns.Normal` centered on the candidate and with a variance equal to SpecFunc.MaxScalar. 34 35The posterior distribution of the parameter is :class:`~openturns.Normal` and reflects the 36variability of the optimum parameter depending on the observation sample. 37The associated covariance matrix may be regularized depending on the value of the 38key `LinearLeastSquaresCalibration-Regularization` in the :class:`~openturns.ResourceMap`. 39Let us denote by :math:`s_1` the smallest singular value of the covariance matrix. 40The default value of the `LinearLeastSquaresCalibration-Regularization`, zero, 41ensures that the singular values of the covariance matrix are left unmodified. 42If this parameter is set to a nonzero, relatively small, value denoted by :math:`\epsilon`, 43then all singular values of the covariance matrix are increased by :math:`\epsilon s_1`. 44 45The resulting distribution of the output error is :class:`~openturns.Normal` with a zero mean 46and a diagonal covariance matrix computed from the residuals. 47The residuals are computed based on the linearization of the model, 48where the Jacobian matrix is evaluated at the candidate point. 49The diagonal of the covariance matrix of the output error 50is constant and is estimated with the unbiased variance estimator. 51 52See also 53-------- 54GaussianLinearCalibration, NonLinearLeastSquaresCalibration, GaussianNonLinearCalibration 55 56Examples 57-------- 58Calibrate a nonlinear model using linear least-squares: 59 60>>> import openturns as ot 61>>> ot.RandomGenerator.SetSeed(0) 62>>> m = 10 63>>> x = [[0.5 + i] for i in range(m)] 64>>> inVars = ['a', 'b', 'c', 'x'] 65>>> formulas = ['a + b * exp(c * x)'] 66>>> model = ot.SymbolicFunction(inVars, formulas) 67>>> p_ref = [2.8, 1.2, 0.5] 68>>> params = [0, 1, 2] 69>>> modelX = ot.ParametricFunction(model, params, p_ref) 70>>> y = modelX(x) 71>>> y += ot.Normal(0.0, 0.05).getSample(m) 72>>> candidate = [1.0]*3 73>>> method = 'SVD' 74>>> algo = ot.LinearLeastSquaresCalibration(modelX, x, y, candidate, method) 75>>> algo.run() 76>>> print(algo.getResult().getParameterMAP()) 77[8.24019,0.0768046,0.992957]" 78 79// --------------------------------------------------------------------- 80 81%feature("docstring") OT::LinearLeastSquaresCalibration::getModelObservations 82"Accessor to the model evaluation at the candidate. 83 84Returns 85------- 86modelObservation : :class:`~openturns.Sample` 87 Evaluation of the model at the candidate point." 88 89// --------------------------------------------------------------------- 90 91%feature("docstring") OT::LinearLeastSquaresCalibration::getGradientObservations 92"Accessor to the model gradient at the candidate. 93 94Returns 95------- 96gradientObservation : :class:`~openturns.Matrix` 97 Gradient of the model at the candidate point." 98 99// --------------------------------------------------------------------- 100 101%feature("docstring") OT::LinearLeastSquaresCalibration::getCandidate 102"Accessor to the parameter candidate. 103 104Returns 105------- 106candidate : :class:`~openturns.Point` 107 Parameter candidate." 108 109// --------------------------------------------------------------------- 110 111%feature("docstring") OT::LinearLeastSquaresCalibration::getMethodName 112"Accessor to the name of least-squares method used for the resolution. 113 114Returns 115------- 116name : :class:`~openturns.String` 117 Name of least-squares method used for the resolution." 118 119