h2o: h2o.splitFrame – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

h2o.splitFrame

Split an H2O Data Set

Description

Split an existing H2O data set according to user-specified ratios. The number of subsets is always 1 more than the number of given ratios. Note that this does not give an exact split. H2O is designed to be efficient on big data using a probabilistic splitting method rather than an exact split. For example, when specifying a split of 0.75/0.25, H2O will produce a test/train split with an expected value of 0.75/0.25 rather than exactly 0.75/0.25. On small datasets, the sizes of the resulting splits will deviate from the expected value more than on big data, where they will be very close to exact.

Usage

h2o.splitFrame(data, ratios = 0.75, destination_frames, seed = -1)

Arguments

`data`	An H2OFrame object, to be split.
`ratios`	A numeric value or array indicating the ratio of total rows contained in each split. Must total up to less than 1. e.g. c(0.8) for 80/20 split.
`destination_frames`	An array of frame IDs equal to the number of values specified in the ratios array, plus one.
`seed`	Random seed.

Value

Returns a list of split H2OFrames

Examples

## Not run: 
library(h2o)
h2o.init()
iris_hf <- as.h2o(iris)
iris_split <- h2o.splitFrame(iris_hf, ratios = c(0.2, 0.5))
head(iris_split[[1]])
summary(iris_split[[1]])

## End(Not run)

h2o

R Interface for the 'H2O' Scalable Machine Learning Platform

v3.32.1.2

Apache License (== 2.0)

Authors

Erin LeDell [aut, cre], Navdeep Gill [aut], Spencer Aiello [aut], Anqi Fu [aut], Arno Candel [aut], Cliff Click [aut], Tom Kraljevic [aut], Tomas Nykodym [aut], Patrick Aboyoun [aut], Michal Kurka [aut], Michal Malohlava [aut], Ludi Rehak [ctb], Eric Eckstrand [ctb], Brandon Hill [ctb], Sebastian Vidrio [ctb], Surekha Jadhawani [ctb], Amy Wang [ctb], Raymond Peck [ctb], Wendy Wong [ctb], Jan Gorecki [ctb], Matt Dowle [ctb], Yuan Tang [ctb], Lauren DiPerna [ctb], Tomas Fryda [ctb], H2O.ai [cph, fnd]

Initial release

2021-04-29