corkscrew (version 1.1)

corkscrew-package: Preprocessor for Data Modeling

Description

Includes binning categorical variables into lesser number of categories based on t-test, converting categorical variables into continuous features using the mean of the response variable for the respective categories, understanding the relationship between the response variable and predictor variables using data transformations.

Details

Package:

corkscrew

Type:

Package

Version:

1.1

Date:

2015-10-30

Depends:

R (>= 3.0.1), ggplot2, gplots, RColorBrewer, igraph, stats, grDevices

License:

GPL (version 2 or newer)

Examples

# using transformation 
data(airquality)
transformation(names(airquality)[2:4],"Ozone",airquality)

# using ctoc
data(ChickWeight)
# Converting the "Chick" variable into factor from ord.factor for demonstration purposes.
ChickWeight$Chick <- as.factor(as.numeric(ChickWeight$Chick))
# Returns a dataframe with two added columns for "Chick" and "Diet"
head(ctoc(y = "weight", x = c("Chick","Diet"), data = ChickWeight, min.obs = 12))

# using tbin
train = as.data.frame(cbind(runif(1000, 10, 1000),sample(1:40, 1000,TRUE)))
colnames(train) = c("response","state")
train$state = as.factor(train$state)
train.output = tbin(dv = "response",idv = c("state"),train,25,TRUE)