public void EhCrossPRESS(int[] group, IMatrix XX, IMatrix YY, IMatrix XU, IMatrix YU)
{
IVector meanX, scaleX, meanY, scaleY;
if (_predictedY == null || _predictedY.Rows != YU.Rows || _predictedY.Columns != YU.Columns)
{
_predictedY = new MatrixMath.BEMatrix(YU.Rows, YU.Columns);
}
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY, out meanX, out scaleX, out meanY, out scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
// allocate the crossPRESS vector here, since now we know about the number of factors a bit more
if (null == _crossPRESS)
{
_crossPRESS = new double[_numFactors + 1]; // one more since we want to have the value at factors=0 (i.e. the variance of the y-matrix)
}
// for all factors do now a prediction of the remaining spectra
_crossPRESS[0] += MatrixMath.SumOfSquares(YU);
for (int nFactor = 1; nFactor <= _numFactors; nFactor++)
{
_analysis.PredictYFromPreprocessed(XU, nFactor, _predictedY);
MultivariateRegression.PostprocessY(_predictedY, meanY, scaleY);
_crossPRESS[nFactor] += MatrixMath.SumOfSquaredDifferences(YU, _predictedY);
}
}
public void EhCrossValidationWorker(int[] group, IMatrix XX, IMatrix YY, IMatrix XU, IMatrix YU)
{
IVector meanX, scaleX, meanY, scaleY;
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY,
out meanX, out scaleX, out meanY, out scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
IROMatrix xResidual = _analysis.SpectralResidualsFromPreprocessed(XU, _numFactors);
if (this._XCrossResiduals == null)
{
this._XCrossResiduals = new MatrixMath.BEMatrix(_numberOfPoints, xResidual.Columns);
}
for (int i = 0; i < group.Length; i++)
{
MatrixMath.SetRow(xResidual, i, this._XCrossResiduals, group[i]);
}
}
public void EhCrossValidationWorker(int[] group, IMatrix XX, IMatrix YY, IMatrix XU, IMatrix YU)
{
IVector meanX, scaleX, meanY, scaleY;
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY,
out meanX, out scaleX, out meanY, out scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
if (_predictedY == null || _predictedY.Rows != YU.Rows || _predictedY.Columns != YU.Columns)
{
_predictedY = new MatrixMath.BEMatrix(YU.Rows, YU.Columns);
}
if (_spectralResidual == null || _spectralResidual.Rows != XU.Rows || _spectralResidual.Columns != _analysis.NumberOfSpectralResiduals)
{
_spectralResidual = new MatrixMath.BEMatrix(XU.Rows, _analysis.NumberOfSpectralResiduals);
}
for (int nFactor = 0; nFactor <= _numFactors; nFactor++)
{
_analysis.PredictedYAndSpectralResidualsFromPreprocessed(XU, nFactor, _predictedY, _spectralResidual);
MultivariateRegression.PostprocessY(_predictedY, meanY, scaleY);
for (int i = 0; i < group.Length; i++)
{
MatrixMath.SetRow(_predictedY, i, _result.GetPredictedYW(NumberOfFactors), group[i]);
MatrixMath.SetRow(_spectralResidual, i, _result.GetSpectralResidualW(NumberOfFactors), group[i]);
}
}
}
public CrossPRESSEvaluator(
int[] spectralRegions,
int numFactors,
ICrossValidationGroupingStrategy groupingStrategy,
SpectralPreprocessingOptions preprocessOptions,
MultivariateRegression analysis
)
: base(spectralRegions, numFactors, groupingStrategy, preprocessOptions, analysis)
{
}
public CrossPredictedXResidualsEvaluator(
int numberOfPoints,
int[] spectralRegions,
int numFactors,
ICrossValidationGroupingStrategy groupingStrategy,
SpectralPreprocessingOptions preprocessOptions,
MultivariateRegression analysis
)
: base(spectralRegions, numFactors, groupingStrategy, preprocessOptions, analysis)
{
_numberOfPoints = numberOfPoints;
}
public CrossPredictedYEvaluator(
int[] spectralRegions,
int numFactors,
ICrossValidationGroupingStrategy groupingStrategy,
SpectralPreprocessingOptions preprocessOptions,
MultivariateRegression analysis,
IMatrix YCrossValidationPrediction
)
: base(spectralRegions, numFactors, groupingStrategy, preprocessOptions, analysis)
{
_YCrossValidationPrediction = YCrossValidationPrediction;
}
public CrossValidationWorker(
int[] spectralRegions,
int numFactors,
ICrossValidationGroupingStrategy groupingStrategy,
SpectralPreprocessingOptions preprocessOptions,
MultivariateRegression analysis
)
{
_spectralRegions = spectralRegions;
_numFactors = numFactors;
_groupingStrategy = groupingStrategy;
_preprocessOptions = preprocessOptions;
_analysis = analysis;
}
public void EhCrossValidationWorker(int[] group, IMatrix <double> XX, IMatrix <double> YY, IMatrix <double> XU, IMatrix <double> YU)
{
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY,
out var meanX, out var scaleX, out var meanY, out var scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
IROMatrix <double> xResidual = _analysis.SpectralResidualsFromPreprocessed(XU, _numFactors);
if (_XCrossResiduals == null)
{
_XCrossResiduals = new MatrixMath.LeftSpineJaggedArrayMatrix <double>(_numberOfPoints, xResidual.ColumnCount);
}
for (int i = 0; i < group.Length; i++)
{
MatrixMath.SetRow(xResidual, i, _XCrossResiduals, group[i]);
}
}
public void EhYCrossPredicted(int[] group, IMatrix <double> XX, IMatrix <double> YY, IMatrix <double> XU, IMatrix <double> YU)
{
if (_predictedY == null || _predictedY.RowCount != YU.RowCount || _predictedY.ColumnCount != YU.ColumnCount)
{
_predictedY = new MatrixMath.LeftSpineJaggedArrayMatrix <double>(YU.RowCount, YU.ColumnCount);
}
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY,
out var meanX, out var scaleX, out var meanY, out var scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
_analysis.PredictYFromPreprocessed(XU, _numFactors, _predictedY);
MultivariateRegression.PostprocessY(_predictedY, meanY, scaleY);
for (int i = 0; i < group.Length; i++)
{
MatrixMath.SetRow(_predictedY, i, _YCrossValidationPrediction, group[i]);
}
}
public void EhYCrossPredicted(int[] group, IMatrix XX, IMatrix YY, IMatrix XU, IMatrix YU)
{
IVector meanX, scaleX, meanY, scaleY;
if (_predictedY == null || _predictedY.Rows != YU.Rows || _predictedY.Columns != YU.Columns)
{
_predictedY = new MatrixMath.BEMatrix(YU.Rows, YU.Columns);
}
MultivariateRegression.PreprocessForAnalysis(_preprocessOptions, _spectralRegions, XX, YY,
out meanX, out scaleX, out meanY, out scaleY);
_analysis.AnalyzeFromPreprocessed(XX, YY, _numFactors);
_numFactors = Math.Min(_numFactors, _analysis.NumberOfFactors);
MultivariateRegression.PreprocessSpectraForPrediction(_preprocessOptions, XU, meanX, scaleX);
_analysis.PredictYFromPreprocessed(XU, _numFactors, _predictedY);
MultivariateRegression.PostprocessY(_predictedY, meanY, scaleY);
for (int i = 0; i < group.Length; i++)
{
MatrixMath.SetRow(_predictedY, i, _YCrossValidationPrediction, group[i]);
}
}
