H2O Generalized Low Rank Models

Generalized Low Rank Models (GLRM) is an algorithm for dimensionality reduction of a dataset

Input

It takes in a DataFrame as input

Type

ml-estimator

Class

fire.nodes.h2o.NodeH2OGlrm

Fields

Name	Title	Description
transform	Transform	STransformation of training data
k	K	Rank of matrix approximation.
ignoredCols	Ignored Columns	Features to be ignored for Modelling
columnsToCategorical	Columns to Categorical	Columns to be Categorical encoded
loss	Loss	Numeric loss function
multiLoss	Multi Loss	Categorical loss function
period	Period	SLength of period (only used with periodic loss function).
regularizationX	Regularization X	Regularization function for X matrix
regularizationY	Regularization Y	Regularization function for X matrix
gammaX	Gamma X	Regularization weight on X matrix
gammaY	Gamma Y	Regularization weight on Y matrix.
maxIterations	Max Iterations	Maximum number of iterations.
maxUpdates	Max Updates	Maximum number of Updated.
initStepSize	Init Step Size	Initial step size.
minStepSize	Min Step Size	Minimum step size.
seed	Seed	Seed for pseudo random number generator (if applicable).
init	Init	STransformation of training data
svdMethod	SVD Method	STransformation of training data
loadingName	Loading Name	[Deprecated] Use representation_name instead. Frame key to save resulting X.
representationName	Representation Name	SFrame key to save resulting X.
expandUserY	Expand User Y	Expand categorical columns in user-specified initial Y.
imputeOriginal	Impute Original	Reconstruct original training data by reversing transform.
recoverSvd	Recover SVD	Recover singular values and eigenvectors of XY.
ignoreConstCols	Ignore Constant Columns	Ignore constant columns.
scoreEachIteration	Score Each Iteration	Whether to score during each iteration of model training.
maxRuntimeSecs	Max Runtime Secs	his argument specifies the maximum time that the AutoML process will run for. If both max_runtime_secs and max_models are specified, then the AutoML run will stop as soon as it hits either of these limits. If neither max_runtime_secs nor max_models are specified, then max_runtime_secs defaults to 3600 seconds (1 hour).
inputCols	Input Columns	Input Columns.
reconstructedCol	Reconstructed Column	Reconstructed Column Name
maxScoringIterations	Max Scoring Iterations	Max Scoring Iterations.

Details

Generalized Low Rank Models (GLRM) is an algorithm for dimensionality reduction of a dataset. It is a general, parallelized optimization algorithm that applies to a variety of loss and regularization functions. Categorical columns are handled by expansion into 0/1 indicator columns for each level. With this approach, GLRM is useful for reconstructing missing values and identifying important features in heterogeneous data.

More details are available at : http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/glrm.html