Data exploration with information theory (weight-of-evidence and information value)

Description

The information package performs exploratory data analysis and variable screening for binary classification models using information theory (WOE and IV).

The package also supports exploratory analysis and variable screening for uplift models (NWOE and NIV).

Note that the only functions you will need to use are create_infotables() and plot_infotables():

- create_infotables() creates WOE or NWOE tables and outputs a variable-strength summary data.frame (IV or NIV)

- plot_infotables() creates WOE or NWOE bar charts for one or more variables

Details

Given a data.frame with a set of predictive variables and a binary response variable, create_infotables() will cycle through all variables and create NWOE or WOE tables. It will also rank all variables by their respective IV or NIV values and return the results in a data.frame.

The package needs minimal inputs. You do not have to explicitly specify which variables to evaluate or provide bins: create_infotables() will process all variables in the dataset and generate appropriate bins for WOE/NWOE analysis.

If requested, calculations can be distributed across multiple cores for better performance.

Note that NWOE analysis is only for uplift models. Thus, for NWOE analysis, you must have a "treatment" and a "control" group in your dataset. The treatment and control groups should identified by a binary indicator variable (1/0).

For regular WOE analysis, on the other hand, all you need is a binary response variable (dependent variable).

You can cross validate your IV or NIV values by supplying a validation dataset. This will produce penalized IV/NIV values.

#' To learn more about the Information package, start with the vignette: browseVignettes(package = "Information")

Author(s)

Kim Larsen (kblarsen4 at gmail.com)

Examples

##------------------------------------------------------------
## WOE analysis, no validation
##------------------------------------------------------------
library(Information)

data(train, package="Information")
train <- subset(train, TREATMENT==1)
IV <- Information::create_infotables(data=train, y="PURCHASE", parallel=FALSE)

print(head(IV$Summary), row.names=FALSE)
print(IV$Tables$N_OPEN_REV_ACTS, row.names=FALSE)

# Plotting a single variable
Information::plot_infotables(IV, "N_OPEN_REV_ACTS")

# Plotting multiple variables
Information::plot_infotables(IV, IV$Summary$Variable[1:4], same_scale=TRUE)

# If the goal is to plot multiple variables individually, as opposed to a comparison-grid, we can
# loop through the variable names and create individual plots
## Not run: 
names <- names(IV$Tables)
plots <- list()
for (i in 1:length(names)){
  plots[[i]] <- plot_infotables(IV, names[i])
}
# Showing the top 18 variables
plots[1:18]

## End(Not run)

# We can speed up create_infotables() by setting parallel=TRUE (default setting)
# If we leave ncore as the default, ncore is set to available clusters - 1
## Not run: 
train <- subset(train, TREATMENT==1)
IV <- Information::create_infotables(data=train, y="PURCHASE")

## End(Not run)
closeAllConnections()

(helper function )Aggregate data for WOE/NWOE calculations

Description

Aggregate returns aggregated data for the WOE and NWOE functions

Usage

Aggregate(data, x, y, breaks, trt)

Arguments

(helper function) Check user inputs and convert factors to characters. Provide readable reasons if errors are found.

Description

CheckInputs Checks user inputs and converts factors to characters. Returns the altered dataframes as a list. Provides readable reasons if errors are found.

Usage

CheckInputs(train, valid, trt, y)

Arguments

(helper function) Plot mutiple WOE vectors on one page

Description

MultiPlot creates a multiple WOE or NWOE bar charts on the same page for a specified vector of variables.

Usage

MultiPlot(information_object, variables, same_scales = FALSE)

Arguments

Create WOE table (helper function)

Description

WOE returns NWOE tables from a data.frame prepared by Information::Aggregate(). This is only for net lift models.

Usage

NWOE(t, x)

Arguments

(helper function) Plot a WOE or NWOE vector

Description

SinglePlot creates a bar chart of the WOE or NWOE pattern for a specified variable

Usage

SinglePlot(information_object, variable, show_values = FALSE)

Arguments

Create WOE tables from aggregated data (helper function)

Description

WOE returns WOE tables from data frames prepared by Aggregate()

Usage

WOE(t, x)

Arguments

Create WOE/NWOE tables and rank variables by IV/NIV

Description

create_infotables returns WOE or NWOE tables (as data.frames), and a data.frame with IV or NIV values for all predictive variables.

Usage

create_infotables(data = NULL, valid = NULL, y = NULL, bins = 10,
  trt = NULL, ncore = NULL, parallel = TRUE)

Arguments

Examples

 
##------------------------------------------------------------
## WOE analysis, no cross validation
##------------------------------------------------------------
library(Information)
data(train, package="Information")
train <- subset(train, TREATMENT==1)
IV <- Information::create_infotables(data=train, y="PURCHASE", parallel=FALSE)
print(head(IV$Summary), row.names=FALSE)
print(IV$Tables$N_OPEN_REV_ACTS, row.names=FALSE)
closeAllConnections()

(helper function) Calculate cross validation penalty

Description

is.binary returns TRUE if a variable is binary, and FALSE otherwise

Usage

is.binary(x)

Arguments

(helper function) Calculate cross validation penalty

Description

penalty returns the weighted cross validation penalty.

Usage

penalty(t, v, d_net_lift)

Arguments

Create bar charts for WOE or NWOE vectors

Description

plot_infotable creates WOE or NWOE bar charts for one or more variables. For multi-variable plots, bar charts will be displayed in a grid for comparison.

Usage

plot_infotables(information_object = NULL, variables = NULL,
  same_scales = FALSE, show_values = FALSE)

Arguments

Examples

 
##------------------------------------------------------------
## WOE plots
##------------------------------------------------------------
library(Information)
data(train, package="Information")
train <- subset(train, TREATMENT==1)
IV <- Information::create_infotables(data=train, y="PURCHASE", parallel=FALSE)

# Plotting a single variable
Information::plot_infotables(IV, "N_OPEN_REV_ACTS")

# Plotting multiple variables in a grid for comparison
Information::plot_infotables(IV, IV$Summary$Variable[1:4], same_scale=TRUE)

# If the goal is to plot multiple variables individually, as opposed to a comparison-grid, we can
# loop through the variable names and create individual plots
## Not run: 
names <- names(IV$Tables)
plots <- list()
for (i in 1:length(names)){
  plots[[i]] <- plot_infotables(IV, names[i])
}
# Showing the top 18 variables
plots[1:18]

## End(Not run)

# We can speed up the creation of the information tables by invoking the parallel option (default)
# If we leave ncore as the default, create_infotables() will set ncore to available clusters - 1
## Not run: 
train <- subset(train, TREATMENT==1)
IV <- Information::create_infotables(data=train, y="PURCHASE")

## End(Not run)
closeAllConnections()

Training dataset

Description

The data is from a historical marketing campaign. It contains 68 predictive variables and 4,972 records. In addition, it contains a treatment indicator and a purchase indicator.

Usage

train

Format

A data frame with 10000 rows and 70 variables:

• TREATMENT: equals 1 if the person received the marketing offer, and 0 if the person was in the control group

• PURCHASE: equals 1 if the person accepted the offer, and 0 otherwise

• UNIQUE_ID: unique identifier

• AGE: age of the person

• D_REGION_X: 1 if the person lives in region X, 0 otherwise (3 regions: A, B, C)

• Other variables are from credit bureau data (e.g., N_OPEN_REV_ACTS = number of open revolving accounts)

Validation dataset

Description

The data is from a historical marketing campaign. It contains 68 predictive variables and 5,060 records. In addition, it contains a treatment indicator and a purchase indicator.

Usage

valid

Format

A data frame with 10000 rows and 70 variables:

• TREATMENT: equals 1 if the person received the marketing offer, and 0 if the person was in the control group

• PURCHASE: equals 1 if the person accepted the offer, and 0 otherwise

• UNIQUE_ID: unique identifier

• AGE: age of the person

• D_REGION_X: 1 if the person lives in region X, 0 otherwise (3 regions: A, B, C)

• Other variables are from credit bureau data (e.g., N_OPEN_REV_ACTS = number of open revolving accounts)