hex.genmodel.algos.glrm

Class GlrmMojoModel

java.lang.Object

hex.genmodel.GenModel

hex.genmodel.MojoModel

hex.genmodel.algos.glrm.GlrmMojoModel

All Implemented Interfaces:

IGenModel, java.io.Serializable, IGeneratedModel

public class GlrmMojoModel
extends MojoModel

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
double	_accuracyEps
double[]	_allAlphas
double[][]	_archetypes
double[][]	_archetypes_raw
int[]	_catOffsets
double	_gammax
GlrmInitialization	_init
int	_iterNumber
GlrmLoss[]	_losses
int	_ncats
int	_ncolA
int	_ncolX
int	_ncolY
int	_nnums
double[]	_normMul
double[]	_normSub
int	_nrowY
int	_numAlphaFactors
int[]	_numLevels
int[]	_permutation
long	_rcnt
GlrmRegularizer	_regx
boolean	_reverse_transform
long	_seed
boolean	_transposed

Fields inherited from class hex.genmodel.MojoModel

_algoName, _balanceClasses, _category, _defaultThreshold, _h2oVersion, _modelAttributes, _modelClassDistrib, _modelDescriptor, _mojo_version, _nclasses, _nfeatures, _priorClassDistrib, _reproducibilityInformation, _supervised, _uuid

Fields inherited from class hex.genmodel.GenModel

_domains, _foldColumn, _names, _offsetColumn, _responseColumn, _treatmentColumn

Constructor Summary

Constructors

Constructor and Description
GlrmMojoModel(java.lang.String[] columns, java.lang.String[][] domains, java.lang.String responseColumn)

Method Summary

Modifier and Type	Method and Description
double	applyBestAlpha(double[] u, double[] x, double[] grad, double[] a, double oldObj, java.util.Random random) This method will try a bunch of arbitray alpha values and pick the best to return which get the best obj improvement.
static int	getCatCidx(int j, int level, int[] numLevels, int[] catOffsets)
java.util.EnumSet<ModelCategory>	getModelCategories() Override this for models that may produce results in different categories.
static int	getNumCidx(int j, int[] catOffsets)
java.lang.String[]	getOutputNames()
int	getPredsSize(ModelCategory mc)
double[]	getRowData(double[] row)
static double[]	impute_data(double[] xfactor, double[] preds, int nnums, int ncats, int[] permutation, boolean reverse_transform, double[] normMul, double[] normSub, GlrmLoss[] losses, boolean transposed, double[][] archetypes_raw, int[] catOffsets, int[] numLevels)
static double[]	initializeAlphas(int numAlpha)
static double[]	lmulCatBlock(double[] x, int j, int[] numLevels, boolean transposed, double[][] archetypes_raw, int[] catOffsets)
static double	lmulNumCol(double[] x, int j, boolean transposed, double[][] archetypes_raw, int[] catOffsets)
static int	rank(boolean transposed, double[][] archetypes_raw)
double[]	score0(double[] row, double[] preds) This function corresponds to the DimReduction model category
double[]	score0(double[] row, double[] preds, long seedValue)

Methods inherited from class hex.genmodel.MojoModel

getModelCategory, getUUID, isSupervised, load, load, load, nclasses, nfeatures

Methods inherited from class hex.genmodel.GenModel

bitSetContains, bitSetIsInRange, calibrateClassProbabilities, convertDouble2Float, correctProbabilities, createAuxKey, features, GBM_rescale, getCategoricalEncoding, getColIdx, getDomainValues, getDomainValues, getDomainValues, getHeader, getNames, getNumClasses, getNumCols, getNumResponseClasses, getOffsetName, getOrigDomainValues, getOrigNames, getOrigNumCols, getOrigProjectionArray, getOutputDomains, getPrediction, getPredictionBinomial, getPredictionMultinomial, getPredsSize, getResponseIdx, getResponseName, GLM_identityInv, GLM_inverseInv, GLM_logInv, GLM_logitInv, GLM_ologitInv, GLM_tweedieInv, img2pixels, internal_threadSafeInstance, isAutoEncoder, isClassifier, KMeans_closest, KMeans_distance, KMeans_distance, KMeans_distances, Kmeans_preprocessData, Kmeans_preprocessData, KMeans_simplex, log_rescale, mapEnum, nCatFeatures, requiresOffset, score0, setCats, setCats, setInput, setInput

Methods inherited from class java.lang.Object

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

Field Detail

_ncolA

public int _ncolA

_ncolX

public int _ncolX

_ncolY

public int _ncolY

_nrowY

public int _nrowY

_archetypes

public double[][] _archetypes

_archetypes_raw

public double[][] _archetypes_raw

_numLevels

public int[] _numLevels

_catOffsets

public int[] _catOffsets

_permutation

public int[] _permutation

_losses

public GlrmLoss[] _losses

_regx

public GlrmRegularizer _regx

_gammax

public double _gammax

_init

public GlrmInitialization _init

_ncats

public int _ncats

_nnums

public int _nnums

_normSub

public double[] _normSub

_normMul

public double[] _normMul

_seed

public long _seed

_transposed

public boolean _transposed

_reverse_transform

public boolean _reverse_transform

_accuracyEps

public double _accuracyEps

_iterNumber

public int _iterNumber

_rcnt

public long _rcnt

_numAlphaFactors

public int _numAlphaFactors

_allAlphas

public double[] _allAlphas

Constructor Detail

GlrmMojoModel

public GlrmMojoModel(java.lang.String[] columns,
                     java.lang.String[][] domains,
                     java.lang.String responseColumn)

Method Detail

getModelCategories

public java.util.EnumSet<ModelCategory> getModelCategories()

Description copied from class: GenModel

Override this for models that may produce results in different categories.

Specified by:

getModelCategories in interface IGenModel

Overrides:

getModelCategories in class GenModel

getPredsSize

public int getPredsSize(ModelCategory mc)

Overrides:

getPredsSize in class GenModel

initializeAlphas

public static double[] initializeAlphas(int numAlpha)

score0

public double[] score0(double[] row,
                       double[] preds,
                       long seedValue)

getRowData

public double[] getRowData(double[] row)

applyBestAlpha

public double applyBestAlpha(double[] u,
                             double[] x,
                             double[] grad,
                             double[] a,
                             double oldObj,
                             java.util.Random random)

This method will try a bunch of arbitray alpha values and pick the best to return which get the best obj improvement.

Parameters:

u -

x -

grad -

a -

oldObj -

random -

Returns:

score0

public double[] score0(double[] row,
                       double[] preds)

This function corresponds to the DimReduction model category

Specified by:

score0 in class GenModel

impute_data

public static double[] impute_data(double[] xfactor,
                                   double[] preds,
                                   int nnums,
                                   int ncats,
                                   int[] permutation,
                                   boolean reverse_transform,
                                   double[] normMul,
                                   double[] normSub,
                                   GlrmLoss[] losses,
                                   boolean transposed,
                                   double[][] archetypes_raw,
                                   int[] catOffsets,
                                   int[] numLevels)

getNumCidx

public static int getNumCidx(int j,
                             int[] catOffsets)

lmulNumCol

public static double lmulNumCol(double[] x,
                                int j,
                                boolean transposed,
                                double[][] archetypes_raw,
                                int[] catOffsets)

getCatCidx

public static int getCatCidx(int j,
                             int level,
                             int[] numLevels,
                             int[] catOffsets)

lmulCatBlock

public static double[] lmulCatBlock(double[] x,
                                    int j,
                                    int[] numLevels,
                                    boolean transposed,
                                    double[][] archetypes_raw,
                                    int[] catOffsets)

rank

public static int rank(boolean transposed,
                       double[][] archetypes_raw)

getOutputNames

public java.lang.String[] getOutputNames()

Overrides:

getOutputNames in class GenModel