From c8e34571a69172803f7b76d64d80357f53395bc1 Mon Sep 17 00:00:00 2001
From: faridm69 <faridmohammadi69@gmail.com>
Date: Thu, 1 Oct 2020 11:41:20 +0200
Subject: [PATCH] [surrogate] fixed bugs.
---
.../BayesInference/BayesInference.py | 4 +--
.../PostProcessing/PostProcessing.py | 2 +-
BayesValidRox/surrogate_models/ExpDesigns.py | 3 +-
.../surrogate_models/RegressionFastARD.py | 1 -
.../surrogate_models/surrogate_models.py | 31 ++++++++++---------
5 files changed, 21 insertions(+), 20 deletions(-)
diff --git a/BayesValidRox/BayesInference/BayesInference.py b/BayesValidRox/BayesInference/BayesInference.py
index 0d7fc81c2..718276889 100644
--- a/BayesValidRox/BayesInference/BayesInference.py
+++ b/BayesValidRox/BayesInference/BayesInference.py
@@ -158,7 +158,7 @@ class BayesInference:
TotalOutputs = np.empty((SampleSize,0))
for outputType in OutputType:
TotalOutputs = np.hstack((TotalOutputs, Outputs[outputType]))
-
+
# Add the std of the PCE is chosen as emulator.
if self.emulator:
covMatrix_PCE = np.zeros((NofMeasurements, NofMeasurements), float)
@@ -178,7 +178,7 @@ class BayesInference:
logL = multivariate_normal.logpdf(TotalOutputs, mean=Data, cov=covMatrix)
except:
logL = -np.inf
-
+
return logL
#--------------------------------------------------------------------------------------------------------
diff --git a/BayesValidRox/PostProcessing/PostProcessing.py b/BayesValidRox/PostProcessing/PostProcessing.py
index 334c9ad7d..c50be9a86 100644
--- a/BayesValidRox/PostProcessing/PostProcessing.py
+++ b/BayesValidRox/PostProcessing/PostProcessing.py
@@ -67,7 +67,7 @@ class PostProcessing:
if PCEModel.DimRedMethod.lower() == 'pca':
PCA = PCEModel.pca[Outkey]
- self.PCEMeans[Outkey] = PCA.inverse_transform(PCEMean)
+ self.PCEMeans[Outkey] = PCA.mean_ + np.dot(PCEMean, PCA.components_)
self.PCEStd[Outkey] = np.sqrt(np.dot(PCEVar, PCA.components_**2))
else:
self.PCEMeans[Outkey] = PCEMean
diff --git a/BayesValidRox/surrogate_models/ExpDesigns.py b/BayesValidRox/surrogate_models/ExpDesigns.py
index 1a1ed840f..ccb9d5d7e 100644
--- a/BayesValidRox/surrogate_models/ExpDesigns.py
+++ b/BayesValidRox/surrogate_models/ExpDesigns.py
@@ -31,7 +31,8 @@ class ExpDesigns:
self.NCandidate = 1
self.PostSnapshot = False
self.stepSnapshot = 2
- self.UtilityFunction = 'None'
+ self.ExploitMethod = 'Space-filling'
+ self.UtilityFunction = 'Space-filling'
self.ExploreMethod = 'random'
self.NrofCandGroups = 4
self.nReprications = 1
diff --git a/BayesValidRox/surrogate_models/RegressionFastARD.py b/BayesValidRox/surrogate_models/RegressionFastARD.py
index 880baf488..93819b75a 100755
--- a/BayesValidRox/surrogate_models/RegressionFastARD.py
+++ b/BayesValidRox/surrogate_models/RegressionFastARD.py
@@ -236,7 +236,6 @@ class RegressionFastARD(LinearModel,RegressorMixin):
# in case of almost perfect multicollinearity between some features
# start from feature 0
if np.sum( XXd - X_mean**2 < np.finfo(np.float32).eps ) > 0:#>= 0:
- print("start from feature 0.")
A[0] = np.finfo(np.float16).eps
active[0] = True
diff --git a/BayesValidRox/surrogate_models/surrogate_models.py b/BayesValidRox/surrogate_models/surrogate_models.py
index 636688190..791b6dff8 100644
--- a/BayesValidRox/surrogate_models/surrogate_models.py
+++ b/BayesValidRox/surrogate_models/surrogate_models.py
@@ -684,8 +684,9 @@ class Metamodel:
LCerror = AllLCerror[str(bestIdx)]
degree = DegreeArray[np.nanargmax(scores)]
qnorm = float(qnorm[bestqIdx])
-
- if verbose:
+
+ # ------------------ Summary of results ------------------
+ if self.DisplayFlag:
# Create PSI_Sparse by removing redundent terms
nnz_idx = np.nonzero(coeffs)[0]
BasisIndices_Sparse = PolynomialDegrees[nnz_idx]
@@ -697,8 +698,8 @@ class Metamodel:
round(len(BasisIndices_Sparse)/P, 3)))
print('Final ModLOO error estimate: %.3e'%(1-max(scores)))
print('\n'+'-'*50)
- # ------------------ Summary of results ------------------
- if self.DisplayFlag:
+
+ if verbose:
print('='*50)
print(' '* 10 +' Summary of results ')
print('='*50)
@@ -867,9 +868,10 @@ class Metamodel:
# Transform via Principal Component Analysis
from sklearn.decomposition import IncrementalPCA as sklearnPCA
- # from sklearn.decomposition import TruncatedSVD as TruncatedSVD
+ # from sklearn.decomposition import TruncatedSVD as sklearnPCA
+
n_samples, n_features = Output.shape
- covar_matrix = sklearnPCA(n_components=None)#, svd_solver='full')
+ covar_matrix = sklearnPCA(n_components=None) #, svd_solver='full')
covar_matrix.fit(Output)
var = np.cumsum(np.round(covar_matrix.explained_variance_ratio_, decimals=5)*100)
try:
@@ -878,14 +880,14 @@ class Metamodel:
selected_n_components = min(n_samples, n_features)
nComponents = min(n_samples, n_features, selected_n_components)
-
- pca = sklearnPCA(n_components=nComponents)#, svd_solver='full')
+ pca = sklearnPCA(n_components=nComponents) #, svd_solver='full')
OutputMatrix = pca.fit_transform(Output)
+ # pca.mean_ = np.zeros(Output.shape[1])
return pca, OutputMatrix
#--------------------------------------------------------------------------------------------------------
- def GaussianProcessEmulator(self, X, y, autoSelect=True, varIdx=None):
+ def GaussianProcessEmulator(self, X, y, autoSelect=False, varIdx=None):
from sklearn.gaussian_process import GaussianProcessRegressor
from sklearn.gaussian_process.kernels import (RBF, Matern, RationalQuadratic,
@@ -914,7 +916,7 @@ class Metamodel:
gp = gp[np.argmax(BME)]
else:
- gp = GaussianProcessRegressor(kernel=kernels[-1], n_restarts_optimizer=3,
+ gp = GaussianProcessRegressor(kernel=kernels[1], n_restarts_optimizer=3,
normalize_y=True)
gp.fit(X, y)
@@ -1074,7 +1076,7 @@ class Metamodel:
# prevBasisIndices = self.BasisDict[key]["y_"+str(idx)] if idx != 0 else None
# outDict = self.adaptiveSparseaPCE(CollocationPoints,OutputMatrix[:,idx],
- # idx,prevBasisIndices,verbose=True)
+ # idx,prevBasisIndices)
@@ -2183,7 +2185,7 @@ class Metamodel:
PCEOutputs_mean[:, idx] = y_mean
PCEOutputs_std[:, idx] = y_std
idx += 1
-
+
# TODO: Plot PCE vs Components
# PCA = self.pca[Outkey]
# self.plot_components(PCA, Y[Outkey], PCEOutputs_mean, PCEOutputs_std, plotname='PCA_PCE_'+Outkey)
@@ -2251,13 +2253,12 @@ class Metamodel:
np.fill_diagonal(covMatrix_PCE, varPCE)
# Aggregate the cov matrices
- covMatrix = covMatrix + covMatrix_PCE
+ covMatrix += covMatrix_PCE
# Compute likelihood
self.Likelihoods = stats.multivariate_normal.pdf(TotalOutputs,
mean=TotalData,
- cov=covMatrix)#,
- # allow_singular=True)
+ cov=covMatrix)
return self.Likelihoods
--
GitLab