Editar

Compartilhar via

rxDTree: Parallel External Memory Algorithm for Classification and Regression Trees

  • Artigo

Description

Fit classification and regression trees on an .xdf file or data frame for small or large data using parallel external memory algorithm.

Usage

  rxDTree(formula, data,
      outFile = NULL, outColName = ".rxNode", writeModelVars = FALSE, extraVarsToWrite = NULL, overwrite = FALSE,
      pweights = NULL, fweights = NULL, method = NULL, parms = NULL, cost = NULL,    
      minSplit = max(20, sqrt(numObs)), minBucket = round(minSplit/3), maxDepth = 10, cp = 0,
      maxCompete = 0, maxSurrogate = 0, useSurrogate = 2, surrogateStyle = 0, xVal = 2,
      maxNumBins = NULL, maxUnorderedLevels = 32, removeMissings = FALSE, 
      computeObsNodeId = NULL, useSparseCube = rxGetOption("useSparseCube"), findSplitsInParallel = TRUE, pruneCp = 0,
      rowSelection = NULL, transforms = NULL, transformObjects = NULL, transformFunc = NULL,
      transformVars = NULL, transformPackages = NULL, transformEnvir = NULL,
      blocksPerRead = rxGetOption("blocksPerRead"), reportProgress = rxGetOption("reportProgress"),
      verbose = 0, computeContext = rxGetOption("computeContext"), 
      xdfCompressionLevel = rxGetOption("xdfCompressionLevel"),
        ...  )

Arguments

formula

formula as described in rxFormula. Currently, formula functions are not supported.

data

either a data source object, a character string specifying a .xdf file, or a data frame object.

outFile

either an RxXdfData data source object or a character string specifying the .xdf file for storing the resulting node indices. If NULL, then no node indices are stored to disk. If the input data is a data frame, the node indices are returned automatically. If rowSelection is specified and not NULL, then outFile cannot be the same as the datasince the resulting set of node indices will generally not have the same number of rows as the original data source.

outColName

character string to be used as a column name for the resulting node indices if outFile is not NULL. Note that make.names is used on outColNameto ensure that the column name is valid. If the outFile is an RxOdbcData source, dots are first converted to underscores. Thus, the default outColName becomes "X_rxNode".

writeModelVars

logical value. If TRUE, and the output file is different from the input file, variables in the model will be written to the output file in addition to the node numbers. If variables from the input data set are transformed in the model, the transformed variables will also be written out.

extraVarsToWrite

NULL or character vector of additional variables names from the input data or transforms to include in the outFile. If writeModelVars is TRUE, model variables will be included as well.

overwrite

logical value. If TRUE, an existing outFilewith an existing column named outColName will be overwritten.

pweights

character string specifying the variable of numeric values to use as probability weights for the observations.

fweights

character string specifying the variable of integer values to use as frequency weights for the observations.

method

character string specifying the splitting method. Currently, only "class" or "anova" are supported. The default is "class" if the response is a factor, otherwise "anova".

parms

optional list with components specifying additional parameters for the "class" splitting method, as follows:

  • prior - a vector of prior probabilities. The priors must be positive and sum to 1. The default priors are proportional to the data counts.
  • loss - a loss matrix, which must have zeros on the diagonal and positive off-diagonal elements. By default, the off-diagonal elements are set to 1.
  • split - the splitting index, either gini (the default) or information.
    If parms is specified, any of the components can be specified or omitted. The defaults will be used for missing components.

cost

a vector of non-negative costs, containing one element for each variable in the model. Defaults to one for all variables. When deciding which split to choose, the improvement on splitting on a variable is divided by its cost.

minSplit

the minimum number of observations that must exist in a node before a split is attempted. By default, this is sqrt(num of obs). For non-XDF data sources, as (num of obs) is unknown in advance, it is wisest to specify this argument directly.

minBucket

the minimum number of observations in a terminal node (or leaf). By default, this is minSplit/3.

cp

numeric scalar specifying the complexity parameter. Any split that does not decrease overall lack-of-fit by at least cp is not attempted.

maxCompete

the maximum number of competitor splits retained in the output. These are useful model diagnostics, as they allow you to compare splits in the output with the alternatives.

maxSurrogate

the maximum number of surrogate splits retained in the output. See the Details for a description of how surrogate splits are used in the model fitting. Setting this to 0 can greatly improve the performance of the algorithm; in some cases almost half the computation time is spent in computing surrogate splits.

useSurrogate

an integer specifying how surrogates are to be used in the splitting process:

  • 0 - display-only; observations with a missing value for the primary split variable are not sent further down the tree.
  • 1 - use surrogates,in order, to split observations missing the primary split variable. If all surrogates are missing, the observation is not split.
  • 2 - use surrogates, in order, to split observations missing the primary split variable. If all surrogates are missing or maxSurrogate=0, send the observation in the majority direction.
    The 0 value corresponds to the behavior of the tree function, and 2 (the default) corresponds to the recommendations of Breiman et al.

xVal

the number of cross-validations to be performed along with the model building. Currently, 1:xVal is repeated and used to identify the folds. If not zero, the cptable component of the resulting model will contain both the mean (xerror) and standard deviation (xstd) of the cross-validation errors, which can be used to select the optimal cost-complexity pruning of the fitted tree. Set it to zero if external cross-validation will be used to evaluate the fitted model because a value of k increases the compute time to (k+1)-fold over a value of zero.

surrogateStyle

an integer controlling selection of a best surrogate. The default, 0, instructs the program to use the total number of correct classifications for a potential surrogate, while 1 instructs the program to use the percentage of correct classification over the non-missing values of the surrogate. Thus, 0 penalizes potential surrogates with a large number of missing values.

maxDepth

the maximum depth of any tree node. The computations take much longer at greater depth, so lowering maxDepth can greatly speed up computation time.

maxNumBins

the maximum number of bins to use to cut numeric data. The default is min(1001, max(101, sqrt(num of obs))). For non-XDF data sources, as (num of obs) is unknown in advance, it is wisest to specify this argument directly. If set to 0, unit binning will be used instead of cutting. See the 'Details' section for more information.

maxUnorderedLevels

the maximum number of levels allowed for an unordered factor predictor for multiclass (>2) classification.

removeMissings

logical value. If TRUE, rows with missing values are removed and will not be included in the output data.

computeObsNodeId

logical value or NULL. If TRUE, the tree node IDs for all the observations are computed and returned. If NULL, the IDs are computed for data.frame with less than 1000 observations and are returned as the where component in the fitted rxDTree object.

useSparseCube

logical value. If TRUE, sparse cube is used.

findSplitsInParallel

logical value. If TRUE, optimal splits for each node are determined using parallelization methods; this will typically speed up computation as the number of nodes on the same level is increased. Note that when it is TRUE, the number of nodes being processed in parallel is also printed to the console, interleaved with the number of rows read from the input data set.

pruneCp

Optional complexity parameter for pruning. If pruneCp > 0, prune.rxDTree is called on the completed tree with the specified pruneCp and the pruned tree is returned. This contrasts with the cp parameter that determines which splits are considered in growing the tree. The option pruneCp="auto" causes rxDTree to call the function rxDTreeBestCp on the completed tree, then use its return value as the cp value for prune.rxDTree.

rowSelection

name of a logical variable in the data set (in quotes) or a logical expression using variables in the data set to specify row selection. For example, rowSelection = "old" will use only observations in which the value of the variable old is TRUE. rowSelection = (age > 20) & (age < 65) & (log(income) > 10) will use only observations in which the value of the age variable is between 20 and 65 and the value of the log of the income variable is greater than 10. The row selection is performed after processing any data transformations (see the arguments transforms or transformFunc). As with all expressions, rowSelection can be defined outside of the function call using the expression function.

transforms

an expression of the form list(name = expression, ...)representing the first round of variable transformations. As with all expressions, transforms (or rowSelection) can be defined outside of the function call using the expression function.

transformObjects

a named list containing objects that can be referenced by transforms, transformsFunc, and rowSelection.

transformFunc

variable transformation function. The ".rxSetLowHigh" attribute must be set for transformed variables if they are to be used in formula. See rxTransform for details.

transformVars

character vector of input data set variables needed for the transformation function. See rxTransform for details.

transformPackages

character vector defining additional R packages (outside of those specified in rxGetOption("transformPackages")) to be made available and preloaded for use in variable transformation functions, e.g., those explicitly defined in RevoScaleR functions via their transforms and transformFunc arguments or those defined implicitly via their formula or rowSelection arguments. The transformPackages argument may also be NULL, indicating that no packages outside rxGetOption("transformPackages") will be preloaded.

transformEnvir

user-defined environment to serve as a parent to all environments developed internally and used for variable data transformation. If transformEnvir = NULL, a new "hash" environment with parent baseenv() is used instead.

blocksPerRead

number of blocks to read for each chunk of data read from the data source.

reportProgress

integer value with options:

  • 0: no progress is reported.
  • 1: the number of processed rows is printed and updated.
  • 2: rows processed and timings are reported.
  • 3: rows processed and all timings are reported.

verbose

integer value. If 0, no verbose output is printed during calculations. Integer values from 1 to 2 provide increasing amounts of information are provided.

computeContext

a valid RxComputeContext. The RxSpark and RxHadoopMR compute contexts distribute the computation among the nodes specified by the compute context; for other compute contexts, the computation is distributed if possible on the local computer.

xdfCompressionLevel

integer in the range of -1 to 9 indicating the compression level for the output data if written to an .xdf file. The higher the value, the greater the amount of compression - resulting in smaller files but a longer time to create them. If xdfCompressionLevel is set to 0, there will be no compression and files will be compatible with the 6.0 release of Revolution R Enterprise. If set to -1, a default level of compression will be used.

...

additional arguments to be passed directly to the Microsoft R Services Compute Engine.

Details

rxDTree is a parallel external memory decision tree algorithm targeted for very large data sets.

It is modeled after rpart (Version 4.1-0) and inspired by the algorithm proposed by Yael Ben-Haim and Elad Tom-Tov (2010).

It uses a histogram as the approximate compressed representation of the data and builds the tree in a breadth-first fashion using horizontal parallelism.

maxNumBins specifies the maximum number of bins for the histogram of each continuous independent variable and thus controls the accuracy of the algorithm. Also, rxDTree builds histograms with roughly equal number of observations in each bin and checks only the boundaries of the bins as candidate splits to find the best split. So it is possible that a suboptimal split is chosen if maxNumBins is too small. This may cause the tree to be different from one constructed by a standard algorithm. Increasing maxNumBins allows more accurate results but with increased time and memory usage..

Surrogate splits may be used to assign observations for which the primary split variable is missing. Surrogate splits compare the groups produced by the remaining predictor variables to the groups produced by the primary split variable, and the predictors are ranked by how well their groups match the primary predictor. The best match is used as the surrogate split.

Value

an object of class "rxDTree" representing the fitted tree. It is a list with components similar to those of class "rpart" with the following distinctions:

  • where - A vector of integers indicating the node to which each point is allocated. This information is always returned if the data source is a data frame. If the data source is not a data frame and outFile is specified. i.e., not NULL, the node indices are written/appended to the specified file with a column name as defined by outColName.

For other components, see rpart.object for details.

Note

rxDTree requires multiple passes over the data set and the maximum number of passes can be computed as follows:

  • quantile computation: 1 pass for computing the quantiles for all continuous variables,

  • recursive partition: maxDepth + 1 passes for building the tree on the entire dataset,

  • tree node assignment: 1 pass for computing the id of the leaf node that each observation falls into.

If cross validation is specified (i.e., xVal>0), additional passes will be needed for each fold:

  • recursive partition: maxDepth + 1 passes for building the tree on the other folds,

  • tree node assignment: 1 pass for predicting on the current fold.

resulting in xVal * ((maxDepth + 1) + 1) additional passes for cross validation.

Author(s)

Microsoft Corporation Microsoft Technical Support

References

Breiman, L., Friedman, J. H., Olshen, R. A. and Stone, C. J. (1984) Classification and Regression Trees. Wadsworth.

Therneau, T. M. and Atkinson, E. J. (2011) An Introduction to Recursive Partitioning Using the RPART Routines.

Yael Ben-Haim and Elad Tom-Tov (2010) A streaming parallel decision tree algorithm. Journal of Machine Learning Research 11, 849--872.

See Also

rpart, rpart.control, rpart.object, rxPredict.rxDTree, rxAddInheritance, rxDForestUtils.

Examples


 set.seed(1234)

 # classification
 iris.sub <- c(sample(1:50, 25), sample(51:100, 25), sample(101:150, 25))
 iris.dtree <- rxDTree(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, 
     data = iris[iris.sub, ])
 iris.dtree

 table(rxPredict(iris.dtree, iris[-iris.sub, ], type = "class")[[1]], 
     iris[-iris.sub, "Species"])

 # regression
 infert.nrow <- nrow(infert)
 infert.sub <- sample(infert.nrow, infert.nrow / 2)
 infert.dtree <- rxDTree(case ~ age + parity + education + spontaneous + induced, 
     data = infert[infert.sub, ], cp = 0.01)
 infert.dtree

 hist(rxPredict(infert.dtree, infert[-infert.sub, ])[[1]] - 
     infert[-infert.sub, "case"])

 # use transformations with rxDTree
 range(log(iris$Sepal.Width)) ## find out the overall low/high values of the transformed variable
 #[1] 0.6931472 1.4816045 

 myTransform <- function(dataList) 
 { 
     dataList$log.Sepal.Width <-log(dataList$Sepal.Width) 
     attr(dataList$log.Sepal.Width, ".rxSetLowHigh") <- c(0.6931472, 1.4816045) # set the overall low/high values
     return(dataList) 
 } 

 frm <- Sepal.Length ~ log.Sepal.Width + Petal.Length + Petal.Width + Species 
 model <- rxDTree(frm, iris, transformFunc = myTransform, transformVars = c("Sepal.Width")) 
 stopifnot("log.Sepal.Width" %in% names(model$variable.importance)) 

 # .xdf file
 claimsXdf <- file.path(rxGetOption("sampleDataDir"),"claims.xdf")
 claims.dtree <- rxDTree(cost ~ age + car.age + type,
   data = claimsXdf)
 claimsCp <- rxDTreeBestCp(claims.dtree)
 claims.dtree1 <- prune.rxDTree(claims.dtree, cp=claimsCp)
 claims.dtree2 <- rxDTree(cost ~ age + car.age + type, 
     data = claimsXdf, pruneCp="auto")
 ## claims.dtree1 and claims.dtree2 should differ only in their 
 ## "call" component
 claims.dtree1[[3]] <- claims.dtree2[[3]] <- NULL
 all.equal(claims.dtree1, claims.dtree2)