ubic.basecode.math.linearmodels

Class LeastSquaresFit

java.lang.Object

ubic.basecode.math.linearmodels.LeastSquaresFit

public class LeastSquaresFit
extends Object

For performing "bulk" linear model fits, but also offers simple methods for simple univariate and multivariate regression for a single vector of dependent variables (data). Has support for ebayes-like shrinkage of variance.

Data with missing values is handled but is less memory efficient and somewhat slower. The main cost is that when there are no missing values, a single QR decomposition can be performed.

Author:

paul

Constructor Summary

Constructors

Constructor and Description
LeastSquaresFit(DesignMatrix designMatrix, DoubleMatrix<String,String> data) Preferred interface if you want control over how the design is set up.
LeastSquaresFit(DesignMatrix designMatrix, DoubleMatrix<String,String> data, cern.colt.matrix.DoubleMatrix2D weights) Weighted least squares fit between two matrices
LeastSquaresFit(DesignMatrix designMatrix, cern.colt.matrix.DoubleMatrix2D b, cern.colt.matrix.DoubleMatrix2D weights) Preferred interface for weighted least squares fit between two matrices
LeastSquaresFit(cern.colt.matrix.DoubleMatrix1D vectorA, cern.colt.matrix.DoubleMatrix1D vectorB) Least squares fit between two vectors.
LeastSquaresFit(cern.colt.matrix.DoubleMatrix1D vectorA, cern.colt.matrix.DoubleMatrix1D vectorB, cern.colt.matrix.DoubleMatrix1D weights) Stripped-down interface for simple use.
LeastSquaresFit(cern.colt.matrix.DoubleMatrix2D A, cern.colt.matrix.DoubleMatrix2D b) ANOVA not possible (use the other constructors)
LeastSquaresFit(cern.colt.matrix.DoubleMatrix2D A, cern.colt.matrix.DoubleMatrix2D b, cern.colt.matrix.DoubleMatrix2D weights) Weighted least squares fit between two matrices
LeastSquaresFit(ObjectMatrix<String,String,Object> sampleInfo, cern.colt.matrix.impl.DenseDoubleMatrix2D data)
LeastSquaresFit(ObjectMatrix<String,String,Object> sampleInfo, cern.colt.matrix.impl.DenseDoubleMatrix2D data, boolean interactions)
LeastSquaresFit(ObjectMatrix<String,String,Object> design, DoubleMatrix<String,String> b) NamedMatrix allows easier handling of the results.
LeastSquaresFit(ObjectMatrix<String,String,Object> design, DoubleMatrix<String,String> data, boolean interactions) NamedMatrix allows easier handling of the results.

Method Summary

Modifier and Type	Method and Description
protected List<GenericAnovaResult>	anova() Compute ANOVA based on the model fit (Type I SSQ, sequential) The idea is to add up the sums of squares (and dof) for all parameters associated with a particular factor.
protected void	ebayesUpdate(double d, double v, cern.colt.matrix.DoubleMatrix1D vp) Provide results of limma eBayes algorithm.
cern.colt.matrix.DoubleMatrix2D	getCoefficients() The matrix of coefficients x for Ax = b (parameter estimates).
double	getDfPrior()
cern.colt.matrix.DoubleMatrix2D	getFitted()
int	getResidualDof()
List<Integer>	getResidualDofs()
cern.colt.matrix.DoubleMatrix2D	getResiduals()
cern.colt.matrix.DoubleMatrix2D	getStudentizedResiduals()
cern.colt.matrix.DoubleMatrix1D	getVarPost()
double	getVarPrior()
cern.colt.matrix.DoubleMatrix2D	getWeights()
boolean	isHasBeenShrunken()
boolean	isHasMissing()
List<LinearModelSummary>	summarize()
List<LinearModelSummary>	summarize(boolean anova)
protected LinearModelSummary	summarize(int i) Compute and organize the various summary statistics for a fit.
Map<String,LinearModelSummary>	summarizeByKeys(boolean anova)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

LeastSquaresFit

public LeastSquaresFit(DesignMatrix designMatrix,
                       DoubleMatrix<String,String> data)

Preferred interface if you want control over how the design is set up.

Parameters:

designMatrix -

data -

LeastSquaresFit

public LeastSquaresFit(DesignMatrix designMatrix,
                       DoubleMatrix<String,String> data,
                       cern.colt.matrix.DoubleMatrix2D weights)

Weighted least squares fit between two matrices

Parameters:

designMatrix -

data -

weights - to be used in modifying the influence of the observations in data.

LeastSquaresFit

public LeastSquaresFit(DesignMatrix designMatrix,
                       cern.colt.matrix.DoubleMatrix2D b,
                       cern.colt.matrix.DoubleMatrix2D weights)

Preferred interface for weighted least squares fit between two matrices

Parameters:

designMatrix -

b - the data

weights - to be used in modifying the influence of the observations in vectorB.

LeastSquaresFit

public LeastSquaresFit(cern.colt.matrix.DoubleMatrix1D vectorA,
                       cern.colt.matrix.DoubleMatrix1D vectorB)

Least squares fit between two vectors. Always adds an intercept!

Parameters:

vectorA - Design

vectorB - Data

LeastSquaresFit

public LeastSquaresFit(cern.colt.matrix.DoubleMatrix1D vectorA,
                       cern.colt.matrix.DoubleMatrix1D vectorB,
                       cern.colt.matrix.DoubleMatrix1D weights)

Stripped-down interface for simple use. Least squares fit between two vectors. Always adds an intercept!

Parameters:

vectorA - Design

vectorB - Data

weights - to be used in modifying the influence of the observations in vectorB.

LeastSquaresFit

public LeastSquaresFit(cern.colt.matrix.DoubleMatrix2D A,
                       cern.colt.matrix.DoubleMatrix2D b)

ANOVA not possible (use the other constructors)

Parameters:

A - Design matrix, which will be used directly in least squares regression

b - Data matrix, containing data in rows.

LeastSquaresFit

public LeastSquaresFit(cern.colt.matrix.DoubleMatrix2D A,
                       cern.colt.matrix.DoubleMatrix2D b,
                       cern.colt.matrix.DoubleMatrix2D weights)

Weighted least squares fit between two matrices

Parameters:

A - Design

b - Data

weights - to be used in modifying the influence of the observations in b. If null, will be ignored.

LeastSquaresFit

public LeastSquaresFit(ObjectMatrix<String,String,Object> sampleInfo,
                       cern.colt.matrix.impl.DenseDoubleMatrix2D data)

Parameters:

sampleInfo - information that will be converted to a design matrix; intercept term is added.

data - Data matrix

LeastSquaresFit

public LeastSquaresFit(ObjectMatrix<String,String,Object> sampleInfo,
                       cern.colt.matrix.impl.DenseDoubleMatrix2D data,
                       boolean interactions)

Parameters:

sampleInfo -

data -

interactions - add interaction term (two-way only is supported)

LeastSquaresFit

public LeastSquaresFit(ObjectMatrix<String,String,Object> design,
                       DoubleMatrix<String,String> b)

NamedMatrix allows easier handling of the results.

Parameters:

design - information that will be converted to a design matrix; intercept term is added.

b - Data matrix

LeastSquaresFit

public LeastSquaresFit(ObjectMatrix<String,String,Object> design,
                       DoubleMatrix<String,String> data,
                       boolean interactions)

NamedMatrix allows easier handling of the results.

Parameters:

design - information that will be converted to a design matrix; intercept term is added.

data - Data matrix

Method Detail

getCoefficients

public cern.colt.matrix.DoubleMatrix2D getCoefficients()

The matrix of coefficients x for Ax = b (parameter estimates). Each column represents one fitted model (e.g., one gene); there is a row for each parameter.

Returns:

getDfPrior

public double getDfPrior()

getFitted

public cern.colt.matrix.DoubleMatrix2D getFitted()

getResidualDof

public int getResidualDof()

getResidualDofs

public List<Integer> getResidualDofs()

getResiduals

public cern.colt.matrix.DoubleMatrix2D getResiduals()

getStudentizedResiduals

public cern.colt.matrix.DoubleMatrix2D getStudentizedResiduals()

Returns:

externally studentized residuals (assumes we have only one QR)

getVarPost

public cern.colt.matrix.DoubleMatrix1D getVarPost()

getVarPrior

public double getVarPrior()

getWeights

public cern.colt.matrix.DoubleMatrix2D getWeights()

isHasBeenShrunken

public boolean isHasBeenShrunken()

isHasMissing

public boolean isHasMissing()

summarize

public List<LinearModelSummary> summarize()

Returns:

summaries. ANOVA will not be computed. If ebayesUpdate has been run, variance and degrees of freedom estimated using the limma eBayes algorithm will be used.

summarize

public List<LinearModelSummary> summarize(boolean anova)

Parameters:

anova - if true, ANOVA will be computed

Returns:

summarizeByKeys

public Map<String,LinearModelSummary> summarizeByKeys(boolean anova)

Parameters:

anova - perform ANOVA, otherwise only basic summarization will be done. If ebayesUpdate has been run, variance and degrees of freedom estimated using the limma eBayes algorithm will be used.

Returns:

anova

protected List<GenericAnovaResult> anova()

Compute ANOVA based on the model fit (Type I SSQ, sequential)

The idea is to add up the sums of squares (and dof) for all parameters associated with a particular factor.

This code is more or less ported from R summary.aov.

Returns:

ebayesUpdate

protected void ebayesUpdate(double d,
                            double v,
                            cern.colt.matrix.DoubleMatrix1D vp)

Provide results of limma eBayes algorithm. These will be used next time summarize is called on this.

Parameters:

d - dfPrior

v - varPrior

vp - varPost

summarize

protected LinearModelSummary summarize(int i)

Compute and organize the various summary statistics for a fit.

If ebayes has been run, variance and degrees of freedom estimated using the limma eBayes algorithm will be used.

Does not populate the ANOVA.

Parameters:

i - index of the fit to summarize

Returns: