h2o (version 2.8.4.4)

h2o.interaction: Create interaction terms between categorical features of an H2O Frame

Description

Create N-th order interaction terms between categorical features of an H2O Frame, N=0,1,2,3,...

Usage

h2o.interaction(data, key=NULL, factors, pairwise, max_factors, min_occurrence)

Arguments

data

An H2OParsedData object containing the variables in the model.

key

The unique hex key assigned to the created frame which has one extra column appended.

factors

Factor columns for which interactions are to be computed. If only one is specified, this can be used to reduce the factor levels.

pairwise

Whether to create pairwise quadratic interactions between factors (otherwise create one higher-order interaction). Only applicable if there are 3 or more factors.

max_factors

Max. number of factor levels in pair-wise interaction terms (if enforced, one extra catch-all factor will be made).

min_occurrence

Min. occurrence threshold for factor levels in pair-wise interaction terms.

Value

• Returns an H2O data frame.

Examples

library(h2o)
localH2O = h2o.init()

# Create some random data
myframe = h2o.createFrame(localH2O, 'framekey', rows = 20, cols = 5,
                          seed = -12301283, randomize = TRUE, value = 0,
                          categorical_fraction = 0.8, factors = 10, real_range = 1,
                          integer_fraction = 0.2, integer_range = 10,
                          binary_fraction = 0, binary_ones_fraction = 0.5,
                          missing_fraction = 0.2,
                          response_factors = 1)
myframe[,3] <- as.factor(myframe[,3])
summary(myframe)
head(myframe, 20)

# Create pairwise interactions
pairwise <- h2o.interaction(myframe, key = 'pairwise', factors = list(c(1,2),c(2,3,4)),
                            pairwise=TRUE, max_factors = 10, min_occurrence = 1)
head(pairwise, 20)
levels(pairwise[,2])

# Create 5-th order interaction
higherorder <- h2o.interaction(myframe, key = 'higherorder', factors = c(1,2,3,4,5),
                               pairwise=FALSE, max_factors = 10000, min_occurrence = 1)
head(higherorder, 20)

# Create a categorical variable out of integer column via self-interaction,
# and keep at most 3 factors, and only if they occur at least twice
summary(myframe[,3])
head(myframe[,3], 20)
trim_integer_levels <- h2o.interaction(myframe, key = 'trim_integers', factors = c(3),
                                       pairwise = FALSE, max_factors = 3, min_occurrence = 2)
head(trim_integer_levels, 20)

# Put all together and clean up temporaries
myframe <- cbind(myframe, pairwise, higherorder, trim_integer_levels)
myframe <- h2o.assign(myframe, 'final.key')
h2o.rm(localH2O, grep(pattern = "Last.value", x = h2o.ls(localH2O)$Key, value = TRUE))
myframe
head(myframe,20)
summary(myframe)
h2o.shutdown(localH2O)