Checking the performance of the bins created on test data

Description

This function uses parallel processing to replicate the bins on test data. This can be used to check the stability of the variable.

Usage

binTest(binObj, testDf, y, xVars, nCores = 1)

Arguments

Value

Returns a list containing 2 elements. The first is a data frame called varSummary which contains a summary of the performance of the variables on the test data including their IV value, entropy, flag which indicates if bad rate increases/decreases with variable value, flag to indicate if a monotonic trend is present, number of bins which flip (i.e. do not follow a monotonic trend), number of bins of the variable and a flag to indicate whether it includes pure nodes (node which do not have any defaults). The second element is a data frame called bin which contains details of all the bins of the variables.

Examples

b1 <- getBins(loanData, "bad_flag", c('LTV', 'balance'))
b2 <- binTest(b1, loanData[1:50,], "bad_flag", c('LTV', 'balance'))

Add binned variables to data

Description

This function creates a data frame with binned variables

Usage

createBins(binObj, df, xVars, prefix = "b_")

Arguments

Value

Returns a dataframe which adds the binned variables to the original data frame

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score', 'balance'), minCr=0.8)
loanData <- createBins(b1, loanData, c('age', 'balance'))

Force a numerical variable to follow a monotonically decreasing trend

Description

This function forces a variable to follow a monotonically decreasing trend by grouping bins. In case such a trend can not be forced a message is printed to the console

Usage

forceDecrTrend(binObj, xVars)

Arguments

Value

Returns a list containing 3 objects. Similar to the getBins function

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score'), minCr=0.6, minProp = 0.01)
b1 <- forceDecrTrend(b1, c('score','age'))

Force a numerical variable to follow a monotonically increasing trend

Description

This function forces a variable to follow a monotonically increasing trend by grouping bins. In case such a trend can not be forced a message is printed to the console

Usage

forceIncrTrend(binObj, xVars)

Arguments

Value

Returns a list containing 3 objects. Similar to the getBins function

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score'), minCr=0.6, minProp = 0.01)
b1 <- forceIncrTrend(b1, c('score','age'))

Bins variables to be used in logistic regression

Description

This function uses parallel processing to compute bins for continuous and categorical variables. The splits are computed using the partykit package which uses conditional inferencing trees. Refer to the package documentation for more details. A separate bin is created for NA values. This can be combined using naCombine function. Categorical variables with a maximum of 10 distinct values are supported.

Usage

getBins(df, y, xVars, minProp = 0.03, minCr = 0.9, nCores = 1)

Arguments

Value

Returns a list containing 3 elements. The first is a data frame called varSummary which contains a summary of all the variables along with their IV value, entropy, p value from ctree function in partykit package, flag which indicates if bad rate increases/decreases with variable value, flag to indicate if a monotonic trend is present, number of bins which flip (i.e. do not follow a monotonic trend), number of bins of the variable and a flag to indicate whether it includes pure nodes (node which do not have any defaults). The second element is a data frame called bin which contains details of all the bins of the variables. The third element is a dataframe called err which contains details of all the variables that could not be split and the reason for the same.

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score', 'balance'))

Simulated default data of 100 customers

Description

A dataset containing simulated data about the characteristic of a customer applying for a loan. The dependent variable is "bad_flag" which indicates whether the customer defaults or not.

Usage

loanData

Format

A data frame with 100 rows and 6 variables:

bad_flag

Indicates whether a customer has defaulted or not

age

Age of the customer

LTV

Ratio of amount of loan to amount of collateral

location

Indicates the location of the customer

balance

Account balance

score

credit score of the customer

Split a variable based on specified cuts

Description

This function splits variables based on cuts that have been input manually

Usage

manualSplit(binObj, splitVar, y, splits, df)

Arguments

Value

Returns a list containing 3 objects. Similar to the getBins function

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score', 'balance'), minCr=0.8)
b1 <- manualSplit(b1, 'age', 'bad_flag', c(25,40,55), loanData)

Combine NA bins

Description

This function combines the NA bin with either the bin having the closest bad rate or the average bad rate if the count of observations in NA bin is low

Usage

naCombine(binObj, xVars, cutoffPropn = 0.01)

Arguments

Value

Returns a list containing 3 objects. Similar to the getBins function

Examples

b1 <- getBins(loanData, "bad_flag", c('age', 'score', 'LTV'))
b1 <- naCombine(b1, c('LTV'))