H2O Gradient Boosting Machine

Gradient Boosting Machine (for Regression and Classification) is a forward learning ensemble method

Input

It takes in a DataFrame as input

Type

ml-estimator

Class

fire.nodes.h2o.NodeH2OGBM

Fields

Name	Title	Description
isResponseIsCategorical	Is Response Column Categorical	Specify a response column type(numeric or categorical). Separates the Classification and Regression
labelCol	Label Column	Response variable column.
featuresCols	Feature Columns	Features to be used for Modelling
columnsToCategorical	Columns to Categorical	Columns to be Categorical encoded
seed	Seed	Seed for pseudo random number generator (if applicable).
balanceClasses	Balance Classes	Balance training data class counts via over/under-sampling (for imbalanced data).
splitRatio	Split Ratio	Split Ratio
nfolds	Number of Folds	Number of folds for K-fold cross-validation (0 to disable or >= 2).
ntrees	Number of Trees	Number of trees.
maxDepth	Max Depth	Maximum tree depth (0 for unlimited).
minRows	Min Rows	Fewest allowed (weighted) observations in a leaf.
nbins	Number of Bins	For numerical columns (real/int), build a histogram of (at least) this many bins, then split at the best point.
nbinsTopLevel	Number of bins top level	For numerical columns (real/int), build a histogram of (at most) this many bins at the root level, then decrease by factor of two per level.
nbinsCats	Number of Bins Categoricals	For categorical columns (factors), build a histogram of this many bins, then split at the best point. Higher values can lead to more overfitting.
categoricalEncoding	Categorical Encoding	Specify one of the various encoding schemes for handling categorical features
ignoreConstCols	Ignore Const Columns	Ignore constant columns.
scoreEachIteration	Score Each Iteration	Whether to score during each iteration of model training.
stoppingRounds	Stopping Rounds	Early stopping based on convergence of stopping_metric. Stop if simple moving average of length k of the stopping_metric does not improve for k:=stopping_rounds scoring events (0 to disable).
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).
stoppingMetric	StoppingMetric	Metric to use for early stopping (AUTO: logloss for classification, deviance for regression)
stoppingTolerance	StoppingTolerance	Relative tolerance for metric-based stopping criterion (stop if relative improvement is not at least this much)
gainsliftBins	Gains Lift Bins	Gains/Lift table number of bins. 0 means disabled.. Default value -1 means automatic binning.
withContributions	With Contributions	Enables or disables generating a sub-column of detailedPredictionCol containing Shapley values.
learnRate	Learn Rate	Learning rate (from 0.0 to 1.0).
advanced	Advanced
convertUnknownCategoricalLevelsToNa	Convert Unknown Categorical Levels to NA	If set to ‘true’, the model converts unknown categorical levels to NA during making predictions.
withLeafNodeAssignments	With Node Assignments	Enables or disables computation of leaf node assignments.
withStageResults	With Stage Results	Enables or disables computation of stage results.
learnRateAnnealing	Learn Rate Annealing	Scale the learning rate by this factor after each tree (e.g., 0.99 or 0.999) .
sampleRate	Sample Rate	Row sample rate per tree (from 0.0 to 1.0).
colSampleRate	Column Sample Rate	Column sample rate(from 0.0 to 1.0).
maxAbsLeafnodePred	Max Absolute Leaf Node Prediction	Maximum absolute value of a leaf node prediction.
predNoiseBandwidth	Prediction Noise Bandwidth	Bandwidth (sigma) of Gaussian multiplicative noise ~N(1,sigma) for tree node predictions.
maxAfterBalanceSize	Max After Balance Size	Maximum relative size of the training data after balancing class counts (can be less than 1.0). Requires balance_classes.
maxConfusionMatrixSize	Max Confusion Matrix Size	[Deprecated] Maximum size (# classes) for confusion matrices to be printed in the Logs.
buildTreeOneNode	Build tree one node	Enables to run on a single node
colSampleRatePerTree	Column Sample Rate Per Tree (from 0.0 to 1.0).	Column sample rate per tree (from 0.0 to 1.0).
colSampleRateChangePerLevel	Column Sample Rate change Per Level	Relative change of the column sampling rate for every level (must be > 0.0 and <= 2.0).
scoreTreeInterval	Score Tree Interval	Score the model after every so many trees. Disabled if set to 0.
minSplitImprovement	Minimum Split Improvement
histogramType	Histogram Type	What type of histogram to use for finding optimal split points. Possible values are
calibrateModel	Calibrate Model	Use Platt Scaling to calculate calibrated class probabilities. Calibration can provide more accurate estimates of class probabilities.
checkConstantResponse	Check Constant Response	UCheck if response column is constant. If enabled, then an exception is thrown if the response column is a constant value.If disabled, then model will train regardless of the response column being a constant value or not.
keepCrossValidationModels	Keep Cross Validation Models	Whether to keep the cross-validated models. Keeping cross-validation models may consume significantly more memory in the H2O cluster.
keepCrossValidationPredictions	Keep Cross Validation Predictions	Whether to keep the predictions of the cross-validation predictions. This needs to be set to TRUE if running the same AutoML object for repeated runs because CV predictions are required to build additional Stacked Ensemble models in AutoML.
keepCrossValidationFoldAssignment	Keep Cross Validation Fold Assignment	Whether to keep cross-validation assignments.
tweediePower	Tweedie Power	Tweedie power for Tweedie regression, must be between 1 and 2.
quantileAlpha	Quantile Alhpa	Desired quantile for Quantile regression, must be between 0 and 1.
huberAlpha	Huber Alpha	Desired quantile for Huber/M-regression (threshold between quadratic and linear loss, must be between 0 and 1).
weightCol	Weight Column	Column with observation weights. Giving some observation a weight of zero is equivalent to excluding it from the dataset; giving an observation a relative weight of 2 is equivalent to repeating that row twice. Negative weights are not allowed. Note: Weights are per-row observation weights and do not increase the size of the data frame. This is typically the number of times a row is repeated, but non-integer values are supported as well. During training, rows with higher weights matter more, due to the larger loss function pre-factor. If you set weight = 0 for a row, the returned prediction frame at that row is zero and this is incorrect. To get an accurate prediction, remove all rows with weight == 0.
offsetCol	Offset Column	Offset column. This will be added to the combination of columns before applying the link function.
foldCol	Fold Column	Column with cross-validation fold index assignment per observation.
foldAssignment	Fold Assignment	Cross-validation fold assignment scheme, if fold_column is not specified. The ‘Stratified’ option will stratify the folds based on the response variable, for classification problems.
aucType	AUC Type	Set default multinomial AUC type.
confusionMatrix	Confusion Matrix
output_confusion_matrix_chart	Output Confusion Matrix Chart	whether to display confusion matrix chart.
cm_chart_title	Confusion Matrix Chart Title	Title name to display in Confusion Matrix Chart
cm_chart_description	Confusion Matrix Chart Description	Description to display in Confusion Matrix CHart
confusionMatrixTargetLegend	Confusion Matrix Target Legend	Legend name to display for Target in Confusion Matrix
confusionMatrixPredictedLabelLegend	Confusion Matrix PredictedLabel Legend	Legend name to display for Predicted Label in Confusion Matrix
confusionMatrixCountLegend	Confusion Matrix Count Legend	Legend name to display for Count in Confusion Matrix
path	Save Confusion Matrix Path	Save Confusion Matrix to Path
Description	Confusion Matrix Description
confusionMatrixRowDescription	Confusion Matrix Outcome description	One can provide the business details of the outcome of the confusion matrix rows
ROC Curve	ROC Curve
output_roc_curve	Output ROC Curve	whether to display confusion matrix chart.
roc_title	ROC Curve Chart Title	Title name to display in ROC Curve Chart
roc_description	ROC Curve Chart Description	Add Description for ROC Curve Chart
xlabel	X Label	X label
ylabel	Y Label	Y Label
Grid Search	Grid Search
paramKeys	Param Name	Param Names. eg: maxDepth ,learnRate, nTrees,distribution
paramValues	Param Value	Enter comma separated values.eg: 4,5,6, eg: gaussian, gamma, bernoulli
gridStrategy	Grid Search Strategy	Strategy to use for model hyperparameter search. Cartesian does exhaustive search; RandomDiscrete searches randomly within given time or model limits.
gridMaxModels	Grid Max Models	Maximum number of models to build in the grid search (0 for unlimited).
gridMaxRuntimeSecs	Grid Max Runtime Seconds	Maximum runtime in seconds for the grid search (0 for unlimited).
gridStoppingRounds	Grid Stopping Rounds	Early stopping based on convergence of the metric during grid search (0 to disable).
gridStoppingTolerance	Grid Stopping Tolerance	Tolerance for metric-based stopping criterion during grid search.
gridStoppingMetric	Grid Stopping Metric	Metric to use for early stopping during grid search (AUTO: logloss for classification, deviance for regression).
gridParallelism	Grid Parallelism	Level of parallelism to use when building models in the grid.
gridSelectBestModelBy	Grid Select Best Model By	Metric used to select the best model from the grid.

Details

Gradient Boosting Machine (for Regression and Classification) is a forward learning ensemble method. The guiding heuristic is that good predictive results can be obtained through increasingly refined approximations. H2O’s GBM sequentially builds regression trees on all the features of the dataset in a fully distributed way - each tree is built in parallel.

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