spark.randomForest {SparkR}R Documentation

Random Forest Model for Regression and Classification


spark.randomForest fits a Random Forest Regression model or Classification model on a SparkDataFrame. Users can call summary to get a summary of the fitted Random Forest model, predict to make predictions on new data, and to save/load fitted models. For more details, see Random Forest Regression and Random Forest Classification


spark.randomForest(data, formula, ...)

## S4 method for signature 'SparkDataFrame,formula'
spark.randomForest(data, formula,
  type = c("regression", "classification"), maxDepth = 5, maxBins = 32,
  numTrees = 20, impurity = NULL, featureSubsetStrategy = "auto",
  seed = NULL, subsamplingRate = 1, minInstancesPerNode = 1,
  minInfoGain = 0, checkpointInterval = 10, maxMemoryInMB = 256,
  cacheNodeIds = FALSE, handleInvalid = c("error", "keep", "skip"))

## S4 method for signature 'RandomForestRegressionModel'

## S3 method for class 'summary.RandomForestRegressionModel'
print(x, ...)

## S4 method for signature 'RandomForestClassificationModel'

## S3 method for class 'summary.RandomForestClassificationModel'
print(x, ...)

## S4 method for signature 'RandomForestRegressionModel'
predict(object, newData)

## S4 method for signature 'RandomForestClassificationModel'
predict(object, newData)

## S4 method for signature 'RandomForestRegressionModel,character', path,
  overwrite = FALSE)

## S4 method for signature 'RandomForestClassificationModel,character', path,
  overwrite = FALSE)



a SparkDataFrame for training.


a symbolic description of the model to be fitted. Currently only a few formula operators are supported, including '~', ':', '+', and '-'.


additional arguments passed to the method.


type of model, one of "regression" or "classification", to fit


Maximum depth of the tree (>= 0).


Maximum number of bins used for discretizing continuous features and for choosing how to split on features at each node. More bins give higher granularity. Must be >= 2 and >= number of categories in any categorical feature.


Number of trees to train (>= 1).


Criterion used for information gain calculation. For regression, must be "variance". For classification, must be one of "entropy" and "gini", default is "gini".


The number of features to consider for splits at each tree node. Supported options: "auto", "all", "onethird", "sqrt", "log2", (0.0-1.0], [1-n].


integer seed for random number generation.


Fraction of the training data used for learning each decision tree, in range (0, 1].


Minimum number of instances each child must have after split.


Minimum information gain for a split to be considered at a tree node.


Param for set checkpoint interval (>= 1) or disable checkpoint (-1). Note: this setting will be ignored if the checkpoint directory is not set.


Maximum memory in MB allocated to histogram aggregation.


If FALSE, the algorithm will pass trees to executors to match instances with nodes. If TRUE, the algorithm will cache node IDs for each instance. Caching can speed up training of deeper trees. Users can set how often should the cache be checkpointed or disable it by setting checkpointInterval.


How to handle invalid data (unseen labels or NULL values) in features and label column of string type in classification model. Supported options: "skip" (filter out rows with invalid data), "error" (throw an error), "keep" (put invalid data in a special additional bucket, at index numLabels). Default is "error".


A fitted Random Forest regression model or classification model.


summary object of Random Forest regression model or classification model returned by summary.


a SparkDataFrame for testing.


The directory where the model is saved.


Overwrites or not if the output path already exists. Default is FALSE which means throw exception if the output path exists.


spark.randomForest returns a fitted Random Forest model.

summary returns summary information of the fitted model, which is a list. The list of components includes formula (formula), numFeatures (number of features), features (list of features), featureImportances (feature importances), maxDepth (max depth of trees), numTrees (number of trees), and treeWeights (tree weights).

predict returns a SparkDataFrame containing predicted labeled in a column named "prediction".


spark.randomForest since 2.1.0

summary(RandomForestRegressionModel) since 2.1.0

print.summary.RandomForestRegressionModel since 2.1.0

summary(RandomForestClassificationModel) since 2.1.0

print.summary.RandomForestClassificationModel since 2.1.0

predict(RandomForestRegressionModel) since 2.1.0

predict(RandomForestClassificationModel) since 2.1.0, character) since 2.1.0, character) since 2.1.0


## Not run: 
##D # fit a Random Forest Regression Model
##D df <- createDataFrame(longley)
##D model <- spark.randomForest(df, Employed ~ ., type = "regression", maxDepth = 5, maxBins = 16)
##D # get the summary of the model
##D summary(model)
##D # make predictions
##D predictions <- predict(model, df)
##D # save and load the model
##D path <- "path/to/model"
##D, path)
##D savedModel <-
##D summary(savedModel)
##D # fit a Random Forest Classification Model
##D t <-
##D df <- createDataFrame(t)
##D model <- spark.randomForest(df, Survived ~ Freq + Age, "classification")
## End(Not run)

[Package SparkR version 2.3.0 Index]