| Interface | Description |
|---|---|
| CoefIndices | |
| DispersionTask.ComputeMaxSumSeriesTsk.CalWVdWVd2WV |
This interface is used to calculate one item of the series in log.
|
| Class | Description |
|---|---|
| ComputationState | |
| ComputationState.GLMSubsetGinfo |
This method will grab a subset of the gradient for each multinomial class.
|
| ComputationState.GramGrad | |
| ComputationState.GramXY |
Cached state of COD (with covariate updates) solver.
|
| ConstrainedGLMUtils | |
| ConstrainedGLMUtils.CoefIndices | |
| ConstrainedGLMUtils.ConstraintGLMStates | |
| ConstrainedGLMUtils.ConstraintsDerivatives | |
| ConstrainedGLMUtils.ConstraintsGram | |
| ConstrainedGLMUtils.LinearConstraintConditions | |
| ConstrainedGLMUtils.LinearConstraints | |
| DispersionTask | |
| DispersionTask.ComputeMaxSumSeriesTsk |
This class will compute the following for every row of the dataset:
1.
|
| DispersionTask.ComputeMaxSumSeriesTsk.EvalLogD2WVEnv | |
| DispersionTask.ComputeMaxSumSeriesTsk.EvalLogDWVEnv | |
| DispersionTask.ComputeMaxSumSeriesTsk.EvalLogWVEnv | |
| DispersionTask.ComputeTweedieConstTsk |
Class to pre-calculate constants assocated with the following processes:
1.
|
| DispersionTask.GenPrediction | |
| DispersionUtils | |
| GLM |
Created by tomasnykodym on 8/27/14.
|
| GLM.BetaInfo | |
| GLM.GLMGradientInfo | |
| GLM.GLMGradientSolver |
Gradient and line search computation for L_BFGS and also L_BFGS solver wrapper (for ADMM)
|
| GLM.GramSolver |
Created by tomasnykodym on 3/30/15.
|
| GLM.PlugValuesImputer | |
| GLM.ProximalGradientInfo | |
| GLM.ProximalGradientSolver |
Simple wrapper around ginfo computation, adding proximal penalty
|
| GLMMetricBuilder |
Class for GLMValidation.
|
| GLMModel |
Created by tomasnykodym on 8/27/14.
|
| GLMModel.GLMOutput | |
| GLMModel.GLMParameters | |
| GLMModel.GLMWeights | |
| GLMModel.GLMWeightsFun | |
| GLMModel.RegularizationPath | |
| GLMModel.Submodel | |
| GLMMojoWriter | |
| GLMScore |
Created by tomas on 3/15/16.
|
| GLMScoringInfo | |
| GLMTask |
All GLM related distributed tasks:
YMUTask - computes response means on actual datasets (if some rows are ignored - e.g ignoring rows with NA and/or doing cross-validation)
GLMGradientTask - computes gradient at given Beta, used by L-BFGS, for KKT condition check
GLMLineSearchTask - computes residual deviance(s) at given beta(s), used by line search (both L-BFGS and IRLSM)
GLMIterationTask - used by IRLSM to compute Gram matrix and response t(X) W X, t(X)Wz
|
| GLMTask.ComputeDiTriGammaTsk |
This function will assist in the estimation of dispersion factors using maximum likelihood
|
| GLMTask.ComputeGammaMLSETsk |
This function will assist in the estimation of dispersion factors using maximum likelihood
|
| GLMTask.ComputeSEorDEVIANCETsk | |
| GLMTask.GLMCoordinateDescentTaskSeqIntercept | |
| GLMTask.GLMCoordinateDescentTaskSeqNaive | |
| GLMTask.GLMGaussianGradientTask | |
| GLMTask.GLMGenerateWeightsTask | |
| GLMTask.GLMIterationTask |
One iteration of glm, computes weighted gram matrix and t(x)*y vector and t(y)*y scalar.
|
| GLMTask.GLMIterationTaskMultinomial | |
| GLMTask.GLMMultinomialGradientBaseTask | |
| GLMTask.GLMMultinomialGradientTask | |
| GLMTask.GLMMultinomialUpdate | |
| GLMTask.GLMMultinomialWLSTask | |
| GLMTask.GLMWLSTask | |
| GLMTask.LSTask |
Task to compute t(X) %*% W %*% X and t(X) %*% W %*% y
|
| GLMTask.YMUTask | |
| GLMUtils | |
| RegressionInfluenceDiagnosticsTasks |
Classes defined here implemented the various pieces of regression influence diagnostics described in this doc:
https://github.com/h2oai/h2o-3/issues/7044.
|
| RegressionInfluenceDiagnosticsTasks.ComputeNewBetaVarEstimatedGaussian | |
| RegressionInfluenceDiagnosticsTasks.RegressionInfluenceDiagBinomial | |
| RegressionInfluenceDiagnosticsTasks.RegressionInfluenceDiagGaussian |
generate DFBETAS as in equation 4 of the document.
|
| TweedieEstimator | |
| TweedieMLDispersionOnly |
class to find bounds on infinite series approximation to calculate tweedie dispersion parameter using the
maximum likelihood function in Dunn et.al.
|