biz.sc.mm.dm.algorithm.regression.linear

Interface LinearRegressionModel

All Superinterfaces:

AppliableObject, ApproximationModel, InputDataInfo, MiningModel, PersistentObject, RegressionModel, ScoringModel, StatisticalModel, SupervisedModel, TaskResult, Verifiable

public interface LinearRegressionModel
extends RegressionModel, ApproximationModel

Common interface for linear regression models (linear, weighted, IRLS).

Field Summary

Fields

Modifier and Type	Field and Description
static int	buildEverything construct all model components (fields)
static int	buildNothing don't construct any complex fields (they'll be constructed in subclass)
static int	buildSweepStats construct only sweep matrix and model stats
static int	modelFull full model (intercept, active and obligatory variables)
static int	modelIntcp model containing only intercept
static int	modelOblig model containing intercept and obligatory variables

Fields inherited from interface biz.sc.mm.dm.base.MiningModel

PROP_INDEX_BUILD_DATA_ID, PROP_INDEX_BUILD_VALID_ID, PROP_INDEX_MODEL_STATS_CLASS, SIGNATURE

Method Summary

Methods

Modifier and Type	Method and Description
java.util.ArrayList	getModelStatName() Gets the list of model statistic names.
double	getModelStatValue(int i) Gets the value of a given model statistic.
double	getPredictedTargetValue(double[] x) Gets model-predicted target value for a given observation x.
double[]	getSignatureBeta() Returns estimated parameters of the current model.
LinearRegressionModelStatistics	getStat() Returns the model statistics.
SweepMatrix	getSweepMatrix() Gets underlying SWEEP matrix for the current regression model
boolean[]	getVar() Returns the array of boolean flags indicating variables included in the model.
java.lang.String	getVarName(int i) Gets the name of a given variable statistic.
double[][]	getVarStat() Gets the values of model variable statistics.
java.util.ArrayList	getVarStatName() Gets the list of variable statistic names.
int	intcp() Returns the number of model intercepts.
int	n() Returns the number of observations used.
int	p() Returns the number of model parameters.
int	ptotal() Returns the total number of available parameters (full model case).
boolean	validateModelParameters() Peforms basic validation of the model parameters: checks whether they are finite real numbers and not all zeroes.

Methods inherited from interface biz.sc.mm.dm.supervised.statistical.StatisticalModel

getCoefficient, getCoefficients, getCorrelation, getCovariance, getImportanceStatisticsPValue, getSignatureNames, getVarNo, isActive

Methods inherited from interface biz.sc.mm.dm.base.MiningModel

getBuildDataIdentifier, getExternalSignature, getFunctionSettings, getInternalFunctionSettings, getInternalSignature, getLearningHistory, getModelStatistics, getSignature, getTransformation, getValidationDataIdentifier, setTransformation

Methods inherited from interface biz.sc.mm.dm.base.TaskResult

isLiberalExecution

Methods inherited from interface biz.sc.mm.dm.PersistentObject

getDescription, getName, getObjectIdentifier, getProperties, isEditable, setDescription, setEditable, setName, setObjectIdentifier

Methods inherited from interface biz.sc.mm.dm.base.Verifiable

verify

Methods inherited from interface biz.sc.mm.dm.supervised.statistical.InputDataInfo

getGroupUsage, getInternalFunctionSettings, getUniqueName, getUsage, getVarGroupName

Field Detail

buildEverything

static final int buildEverything

construct all model components (fields)

See Also:

Constant Field Values

buildSweepStats

static final int buildSweepStats

construct only sweep matrix and model stats

See Also:

Constant Field Values

buildNothing

static final int buildNothing

don't construct any complex fields (they'll be constructed in subclass)

See Also:

Constant Field Values

modelIntcp

static final int modelIntcp

model containing only intercept

See Also:

Constant Field Values

modelOblig

static final int modelOblig

model containing intercept and obligatory variables

See Also:

Constant Field Values

modelFull

static final int modelFull

full model (intercept, active and obligatory variables)

See Also:

Constant Field Values

Method Detail

getSignatureBeta

double[] getSignatureBeta()

Returns estimated parameters of the current model.

Returns:

The vector of estimated model parameters.

getVar

boolean[] getVar()

Returns the array of boolean flags indicating variables included in the model.

Returns:

the array of boolean flags indicating variables included in the model.

getStat

LinearRegressionModelStatistics getStat()

Returns the model statistics.

Returns:

the model statistics.

n

int n()

Returns the number of observations used.

Specified by:

n in interface InputDataInfo

Returns:

the number of observations used.

p

int p()

Returns the number of model parameters.

Specified by:

p in interface StatisticalModel

Returns:

the number of model parameters.

ptotal

int ptotal()

Returns the total number of available parameters (full model case).

Specified by:

ptotal in interface InputDataInfo

Returns:

the total number of available parameters (full model case).

intcp

int intcp()

Returns the number of model intercepts.

Specified by:

intcp in interface InputDataInfo

Returns:

the number of model intercepts.

getPredictedTargetValue

double getPredictedTargetValue(double[] x)

Gets model-predicted target value for a given observation x.

Parameters:

x - vector to be scored.

Returns:

the value of linear score for a given vector x.

validateModelParameters

boolean validateModelParameters()

Peforms basic validation of the model parameters: checks whether they are finite real numbers and not all zeroes.

Returns:

true if model seems to be correct.

getModelStatValue

double getModelStatValue(int i)

Gets the value of a given model statistic.

Parameters:

i - identifier of model statistic.

Returns:

the value of a given model statistic.

getVarStatName

java.util.ArrayList getVarStatName()

Gets the list of variable statistic names.

Returns:

the list of variable statistic names.

getModelStatName

java.util.ArrayList getModelStatName()

Gets the list of model statistic names.

Returns:

the list of model statistic names.

getVarStat

double[][] getVarStat()

Gets the values of model variable statistics.

Returns:

values of model variable statistics.

getVarName

java.lang.String getVarName(int i)

Gets the name of a given variable statistic.

Specified by:

getVarName in interface InputDataInfo

Parameters:

i - identifier of the statistic.

Returns:

the name of a given variable statistic.

getSweepMatrix

SweepMatrix getSweepMatrix()

Gets underlying SWEEP matrix for the current regression model