Merge branch 'master' of github.com:0xdata/h2o

R/h2o-package/R/Algorithms.R

@@ -7,22 +7,28 @@ setGeneric("h2o.prcomp", function(data, tol = 0, standardize = TRUE, retx = FALS
 setGeneric("h2o.pcr", function(x, y, data, ncomp, family, nfolds = 10, alpha = 0.5, lambda = 1.0e-5, tweedie.p = ifelse(family=="tweedie", 0, NA)) { standardGeneric("h2o.pcr") })
 setGeneric("h2o.randomForest", function(x, y, data, ntree = 50, depth = 2147483647, classwt = as.numeric(NA)) { standardGeneric("h2o.randomForest") })
 setGeneric("h2o.getTree", function(forest, k, plot = FALSE) { standardGeneric("h2o.getTree") })
-setGeneric("h2o.gbm", function(x, y, data, n.trees = 100, interaction.depth = 5, n.minobsinnode = 10, shrinkage = 0.1) { standardGeneric("h2o.gbm") })
+setGeneric("h2o.gbm", function(x, y, distribution='multinomial', data, n.trees = 10, interaction.depth = 8, n.minobsinnode = 10, shrinkage = 0.2) { standardGeneric("h2o.gbm") })
 # setGeneric("h2o.gbmgrid", function(x, y, data, n.trees = c(10,100), interaction.depth = c(1,5,10), n.minobsinnode = 10, shrinkage = c(0.01,0.1,0.2)) { standardGeneric("h2o.gbmgrid") })
 setGeneric("h2o.predict", function(object, newdata) { standardGeneric("h2o.predict") })
 
 #----------------------- Generalized Boosting Machines (GBM) -----------------------#
-setMethod("h2o.gbm", signature(x="numeric", y="numeric", data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
-   function(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+setMethod("h2o.gbm", signature(x="numeric", y="numeric", distribution='character', data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
+   function(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
       if (length(x) < 1) stop("GBM requires at least one explanatory variable")
       if(any( x < 1 | x > ncol(data))) stop(paste('Out of range explanatory variable', paste(x[which(x < 1 || x > ncol(data))], collapse=',')))
       if( y < 1 || y > ncol(data) ) stop(paste('Response variable index', y, 'is out of range'))
       if( y %in% x ) stop(paste(colnames(data)[y], 'is both an explanatory and dependent variable'))
       x <- x - 1
       cols=paste(x,collapse=',')
+
+      if( missing(distribution) )
+        distribution <- 'multinomial'
+      if( !(distribution %in% c('multinomial', 'gaussian')) )
+        stop(paste(distribution, "is not a valid distribution; only [multinomial, guassian] are supported"))
+      classification <- ifelse(distribution == 'multinomial', 1, ifelse(distribution=='gaussian', 0, -1))
 
       destKey = paste("__GBMModel_", UUIDgenerate(), sep="")
-      res = h2o.__remoteSend(data@h2o, h2o.__PAGE_GBM, destination_key=destKey, source=data@key, response=colnames(data)[y], cols=cols, ntrees=n.trees, max_depth=interaction.depth, learn_rate=shrinkage, min_rows=n.minobsinnode)
+      res = h2o.__remoteSend(data@h2o, h2o.__PAGE_GBM, destination_key=destKey, source=data@key, response=colnames(data)[y], cols=cols, ntrees=n.trees, max_depth=interaction.depth, learn_rate=shrinkage, min_rows=n.minobsinnode, classification=classification)
       while(h2o.__poll(data@h2o, res$job_key) != -1) { Sys.sleep(1) }
       res2 = h2o.__remoteSend(data@h2o, h2o.__PAGE_GBMModelView, '_modelKey'=destKey)
 
@@ -45,25 +51,28 @@ setMethod("h2o.gbm", signature(x="numeric", y="numeric", data="H2OParsedData", n
       new("H2OGBMModel", key=destKey, data=data, model=result)
 	})
 
-setMethod("h2o.gbm", signature(x="numeric", y="character", data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
-    function(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+setMethod("h2o.gbm", signature(x="numeric", y="character", distribution='ANY', data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
+    function(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
       cc <- colnames( data )
       if( !(y %in% cc) ) stop(paste(y, 'is not a valid column name'))
       y_i <- which(y==cc)
-      h2o.gbm(x, y_i, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
+      h2o.gbm(x, y_i, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
   })
 
-setMethod("h2o.gbm", signature(x="character", y="character", data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
-    function(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+setMethod("h2o.gbm", signature(x="character", y="character", distribution='ANY', data="H2OParsedData", n.trees="numeric", interaction.depth="numeric", n.minobsinnode="numeric", shrinkage="numeric"),
+    function(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
       cc <- colnames( data )
       if( y %in% x ) stop(paste(y, 'is both an explanatory and dependent variable'))
       if(any(!(x %in% cc))) stop(paste(paste(x[which(!(x %in% cc))], collapse=','), 'is not a valid column name'))
       x_i = match(x, cc)
-      h2o.gbm(x_i, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
+      h2o.gbm(x_i, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
   })
 
-setMethod("h2o.gbm", signature(x="ANY", y="character", data="H2OParsedData", n.trees="ANY", interaction.depth="ANY", n.minobsinnode="ANY", shrinkage="ANY"),
-   function(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+setMethod("h2o.gbm", signature(x="ANY", y="character", distribution='ANY', data="H2OParsedData", n.trees="ANY", interaction.depth="ANY", n.minobsinnode="ANY", shrinkage="ANY"),
+   function(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+      if( missing(distribution) ) distribution='multinomial'
+      if( !(distribution %in% c('multinomial', 'gaussian')) )
+        stop(paste(distribution, "is not a valid distribution; only [multinomial, guassian] are supported"))
       if(!(missing(x) || class(x) == "numeric" || class(x) == "character"))
          stop(paste("x cannot be of class", class(x)))
       else if(!(missing(n.trees) || class(n.trees) == "numeric"))
@@ -75,13 +84,13 @@ setMethod("h2o.gbm", signature(x="ANY", y="character", data="H2OParsedData", n.t
       else if(!(missing(n.minobsinnode) || class(n.minobsinnode) == "numeric"))
          stop(paste("n.minobsinnode cannot be of class", class(n.minobsinnode)))
       if(missing(x)) x = setdiff(colnames(data), y)
-      h2o.gbm(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
+      h2o.gbm(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
   })
 
-setMethod("h2o.gbm", signature(x="ANY", y="numeric", data="H2OParsedData", n.trees="ANY", interaction.depth="ANY", n.minobsinnode="ANY", shrinkage="ANY"),
-    function(x, y, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
+setMethod("h2o.gbm", signature(x="ANY", y="numeric", distribution='ANY', data="H2OParsedData", n.trees="ANY", interaction.depth="ANY", n.minobsinnode="ANY", shrinkage="ANY"),
+    function(x, y, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage) {
       if( y < 1 || y > ncol( data ) ) stop(paste(y, 'is not a valid column index'))
-      h2o.gbm(x, colnames(data)[y], data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
+      h2o.gbm(x, colnames(data)[y], data, distribution, n.trees, interaction.depth, n.minobsinnode, shrinkage)
     })
 
 #----------------------------- Generalized Linear Models (GLM) ---------------------------#

R/h2o-package/demo/h2o.gbm.R

@@ -1,9 +1,9 @@
 # This is a demo of H2O's GBM function
 # It imports a data set, parses it, and prints a summary
 # Then, it runs GBM on a subset of the dataset
-library(h2o)
-localH2O = new("H2OClient", ip="localhost", port=54321)
-h2o.checkClient(localH2O)
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
 
 prostate.hex = h2o.importFile(localH2O, system.file("extdata", "prostate.csv", package="h2o"), "prostate.hex")
 summary(prostate.hex)

R/h2o-package/demo/h2o.glm.R

@@ -2,9 +2,9 @@
 # It imports a data set, parses it, and prints a summary
 # Then, it runs GLM with a binomial link function using 10-fold cross-validation
 # Note: This demo runs H2O on localhost:54321
-library(h2o)
-localH2O = new("H2OClient", ip = "localhost", port = 54321)
-h2o.checkClient(localH2O)
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
 
 prostate.hex = h2o.importFile(localH2O, path = system.file("extdata", "prostate.csv", package="h2o"), key = "prostate.hex")
 summary(prostate.hex)
@@ -17,4 +17,4 @@ axis(1, at = 1:length(myLabels), labels = myLabels)
 abline(h = 0, col = 2, lty = 2)
 title("Coefficients from Logistic Regression\n of Prostate Cancer Data")
 
-barplot(prostate.glm@model$coefficients, main = "Coefficients from Logistic Regression\n of Prostate Cancer Data")
+barplot(prostate.glm@model$coefficients, main = "Coefficients from Logistic Regression\n of Prostate Cancer Data")

R/h2o-package/demo/h2o.kmeans.R

@@ -2,9 +2,9 @@
 # It imports a data set, parses it, and prints a summary
 # Then, it runs K-Means with k = 5 centers on a subset of characteristics
 # Note: This demo runs H2O on localhost:54321
-library(h2o)
-localH2O = new("H2OClient", ip="localhost", port=54321)
-h2o.checkClient(localH2O)
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
 
 prostate.hex = h2o.importFile(localH2O, system.file("extdata", "prostate.csv", package="h2o"), "prostate.hex")
 summary(prostate.hex)
@@ -34,4 +34,4 @@ par(mfrow = c(1,2))
 prostate.ctrs = as.data.frame(prostate.km@model$centers)
 plot(prostate.ctrs[,1:2])
 plot(prostate.ctrs[,3:4])
-title("K-Means Centers for k = 10", outer = TRUE, line = -2.0)
+title("K-Means Centers for k = 10", outer = TRUE, line = -2.0)

R/h2o-package/demo/h2o.prcomp.R

@@ -1,9 +1,9 @@
 # This is a demo of H2O's PCA function
 # It imports a data set, parses it, and prints a summary
 # Then, it runs PCA on a subset of the features
-library(h2o)
-localH2O = new("H2OClient", ip="localhost", port=54321)
-h2o.checkClient(localH2O)
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
 
 australia.hex = h2o.importFile(localH2O, system.file("extdata", "australia.csv", package="h2o"), "australia.hex")
 summary(australia.hex)
@@ -13,4 +13,4 @@ print(australia.pca)
 plot(australia.pca)
 
 australia.pca2 = h2o.prcomp(australia.hex, tol = 0.5, standardize = FALSE)
-print(australia.pca2)
+print(australia.pca2)

R/h2o-package/demo/h2o.randomForest.R

@@ -2,9 +2,9 @@
 # It imports a data set, parses it, and prints a summary
 # Then, it runs RF with 50 trees, maximum depth of 100, using the iris class as the response
 # Note: This demo runs H2O on localhost:54321
-library(h2o)
-localH2O = new("H2OClient", ip="localhost", port=54321)
-h2o.checkClient(localH2O)
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
 
 iris.hex = h2o.importFile(localH2O, path = system.file("extdata", "iris.csv", package="h2o"), key = "iris.hex")
 summary(iris.hex)
@@ -16,4 +16,4 @@ invisible(readline("Hit <Return> to continue: "))
 covtype.hex = h2o.importFile(localH2O, path = system.file("extdata", "covtype.csv", package="h2o"), key = "covtype.hex")
 summary(covtype.hex)
 covtype.rf = h2o.randomForest(y = "Cover_Type", x = setdiff(colnames(covtype.hex), c("Cover_Type", "Aspect", "Hillshade_9am")), data = covtype.hex, ntree = 50, depth = 150)
-print(covtype.rf)
+print(covtype.rf)

R/h2o-package/man/H2OClient-class.Rd

@@ -49,7 +49,6 @@ Objects can be created by calls of the form \code{new("H2OClient", ...)}
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 showClass("H2OClient")
 }
 \keyword{classes}

R/h2o-package/man/h2o.checkClient.Rd

@@ -44,7 +44,6 @@ h2o.checkClient(object)
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 h2o.checkClient(localH2O)
 }
 % Add one or more standard keywords, see file 'KEYWORDS' in the

R/h2o-package/man/h2o.gbm.Rd

@@ -0,0 +1,77 @@
+\name{h2o.gbm}
+\alias{h2o.gbm}
+%- Also NEED an '\alias' for EACH other topic documented here.
+\title{
+H2O: GBM
+}
+%%  ~~function to do ... ~~
+
+\description{Builds gradient boosed classification trees on a parsed data set.
+}
+\usage{
+h2o.gbm(y, x, distribution, data, n.trees, interaction.depth, n.minobsinnode, shrinkage)
+}
+\arguments{
+  \item{y}{
+The name or index of the response variable. If the data does not contain a header, this is the column index number starting at 0, and increasing from left to right. (The response must be either an integer or a categorical variable).
+}
+  \item{x}{
+A vector containing the names or indices of the predictor variables to use in building the GBM model.
+}
+  \item{distribution}{
+The type of GBM model to be produced, categorization is "multinomial" (default), "gaussian" used for regression.
+}
+  \item{data}{
+An \code{\linkS4class{H2OParsedData}} object containing the variables in the model.
+}
+  \item{n.trees}{
+Number of trees to grow. Must be a nonnegative integer.
+}
+  \item{interaction.depth}{
+Maximum depth to grow the tree.
+}
+  \item{n.minobsinnode}{
+Minimum number of rows to assign to teminal nodes.
+}
+  \item{shrinkage}{
+A learning-rate parameter defining step size reduction.
+}   
+  }
+\value{
+An object of class \code{\linkS4class{H2OGBM}} with slots key, data, and model, where the last is a list of the following components:
+\item{type }{The type of the tree, which currently must be classification.}
+\item{n.trees }{Number of trees grown.}
+\item{oob_err }{Out of bag error rate.}
+\item{forest }{A matrix giving the minimum, mean, and maximum of the tree depth and number of leaves.}
+\item{confusion }{Confusion matrix of the prediction.}
+}
+\references{
+
+1. Elith, Jane, John R Leathwick, and Trevor Hastie. "A Working Guide to
+Boosted Regression Trees." Journal of Animal Ecology 77.4 (2008): 802-813
+
+2. Friedman, Jerome, Trevor Hastie, Saharon Rosset, Robert Tibshirani,
+and Ji Zhu. "Discussion of Boosting Papers." Ann. Statist 32 (2004): 
+102-107
+
+3. Hastie, Trevor, Robert Tibshirani, and J Jerome H Friedman. The
+Elements of Statistical Learning.
+Vol.1. N.p.: Springer New York, 2001. 
+http://www.stanford.edu/~hastie/local.ftp/Springer/OLD//ESLII_print4.pdf
+}
+
+
+\seealso{
+For more information see: http://docs.0xdata.com
+}
+\examples{
+library(h2oWrapper)
+h2oWrapper.installDepPkgs()
+localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
+# Run classification GBM on CAPSULE ~ AGE + RACE + PSA + DCAPS
+prostate.hex = h2o.importURL(localH2O, path = "https://raw.github.com/0xdata/h2o/master/smalldata/logreg/prostate.csv", key = "prostate.hex")
+h2o.gbm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), data = prostate.hex, n.trees = 100, interaction.depth = 8, n.minobsinnode = 10, shrinkage = 0.2)
+# Run regression GBM on CAPSULE ~ AGE + RACE + PSA + DCAPS
+myX = setdiff(colnames(prostate.hex), c("ID", "DPROS", "DCAPS", "VOL"))
+h2o.gbm(y = "VOL", x = myX, data = prostate.hex, n.trees = 10, interaction.depth = 5, shrinkage = 0.1)
+}

R/h2o-package/man/h2o.getTree.Rd

@@ -48,7 +48,6 @@ The particular tree to retrieve. (Must be an integer between 1 and \code{ntree})
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 irisPath = system.file("extdata", "iris.csv", package="h2o")
 iris.hex = h2o.importFile(localH2O, path = irisPath, key = "iris.hex")
 iris.rf = h2o.randomForest(y = 4, data = iris.hex, ntree = 50, depth = 100)

R/h2o-package/man/h2o.glm.Rd

@@ -80,7 +80,6 @@ The slot xval is a list of \code{\linkS4class{H2OGLMModel}} objects representing
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 # Run GLM of CAPSULE ~ AGE + RACE + PSA + DCAPS
 prostate.hex = h2o.importURL(localH2O, path = "https://raw.github.com/0xdata/h2o/master/smalldata/logreg/prostate.csv", key = "prostate.hex")
 h2o.glm(y = "CAPSULE", x = c("AGE","RACE","PSA","DCAPS"), data = prostate.hex, family = 

R/h2o-package/man/h2o.importFile.Rd

@@ -55,7 +55,6 @@ If \code{parse = TRUE}, the function returns an object of class \code{\linkS4cla
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 irisPath = system.file("extdata", "iris.csv", package="h2o")
 iris.hex = h2o.importFile(localH2O, path = irisPath, key = "iris.hex")
 summary(iris.hex)

R/h2o-package/man/h2o.importFolder.Rd

@@ -49,7 +49,6 @@ WARNING: In H2O, import is lazy! Do not modify the data files on hard disk until
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 myPath = paste(path.package("h2o"), "extdata", sep="/")
 all_files.hex = h2o.importFolder(localH2O, path = myPath)
 for(i in 1:length(all_files.hex))

R/h2o-package/man/h2o.importHDFS.Rd

@@ -52,7 +52,6 @@ When the path is a directory, if \code{parse = TRUE}, the function returns a lis
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 covtype.hex = importHDFS(localH2O, path = "hdfs://192.168.1.173:54321/0xdiag/datasets/standard/covtype.data", parse = TRUE)
 summary(covtype.hex)
 

R/h2o-package/man/h2o.importURL.Rd

@@ -55,7 +55,6 @@ If \code{parse = TRUE}, the function returns an object of class \code{\linkS4cla
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 prostate.hex = h2o.importURL(localH2O, path = "https://raw.github.com/0xdata/h2o/master/smalldata/logreg/prostate.csv", key = "prostate.hex")
 summary(prostate.hex)
 }

R/h2o-package/man/h2o.kmeans.Rd

@@ -60,7 +60,6 @@ An object of class \code{\linkS4class{H2OKMeansModel}} with slots key, data, and
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 prosPath = system.file("extdata", "prostate.csv", package="h2o")
 prostate.hex = h2o.importFile(localH2O, path = prosPath)
 h2o.kmeans(data = prostate.hex, centers = 10, cols = c("AGE", "RACE", "VOL", "GLEASON"))

R/h2o-package/man/h2o.parseRaw.Rd

@@ -48,7 +48,6 @@ After the raw data file is parsed, it will be automatically deleted from the H2O
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 prosPath = system.file("extdata", "prostate.csv", package="h2o")
 prostate.raw = h2o.importFile(localH2O, path = prosPath, parse = FALSE)
 # Do not modify prostate.csv on disk at this point!

R/h2o-package/man/h2o.prcomp.Rd

@@ -50,7 +50,6 @@ The signs of the columns of the rotation matrix are arbitrary, and so may differ
 library(h2oWrapper)
 h2oWrapper.installDepPkgs()
 localH2O = h2oWrapper.init(ip = "localhost", port = 54321, startH2O = TRUE, silentUpgrade = TRUE, promptUpgrade = FALSE)
-library(h2o)
 prosPath = system.file("extdata", "prostate.csv", package="h2o")
 prostate.hex = h2o.importFile(localH2O, path = prosPath)
 prostate.pca = h2o.prcomp(data = prostate.hex, standardize = TRUE)