featureComb_fit_transform: featureComb_fit_transform

In 2econsulting/rAutoFE:

Description

featureComb_fit_transform

Usage

featureComb_fit_transform(dt, vi)

Arguments

dt	data.table object

Examples

library(data.table)
library(binr)
library(rAutoFE)
library(rAutoFS)
library(h2o)
churn <- as.data.table(churn)
splits <- rAutoFE::splitFrame(dt=churn, ratio = c(0.5, 0.3), seed = 1234)
train <- splits[[1]]
valid <- splits[[2]]
test  <- splits[[3]]
h2o.init()
data_hex <- as.h2o(rbind(train, valid))
y = "Churn."
x = colnames(data_hex)[colnames(data_hex)!=y]
ml <- autoFS(data_hex, x, y, num_of_model=5, num_of_vi=10)
train <- featureComb_fit_transform(dt=train, vi=ml$top_vi)
valid <- featureComb_fit_transform(dt=valid, vi=ml$top_vi)
test  <- featureComb_fit_transform(dt=test, vi=ml$top_vi)

2econsulting/rAutoFE documentation built on May 24, 2019, 2:49 p.m.