woe_cal_table: woe_cal_table

In nguyenngocbinh/scorecard.sea: An implementation of score card

Description

Compute the woe_cal_table that shows the Weights Of Evidence (WOE) for each group and respeective Information Values (IVs).

Usage

woe_cal_table(df, y, valueOfBad = 1)

Arguments

df	data.frame.
y	The actual 1/0 flags for the binary response variable. It can take values of either 1 or 0, where 1 represents the 'Bad' or 'Events' while 0 represents 'Good' or 'Non-Events'.
valueOfBad	The value in y that is used to represent 'Bad' or the occurence of the event of interest. Defaults to 1.

Details

For a given actual for a Binary Y variable and a data.frame, the WOE table is generated with calculated WOE's and IV's

Value

The WOE table with the respective weights of evidence for each group and the IV's.

• group The groups (levels) of the categorical X variable for which WOE is to be calculated.

• total. The total number of observations in respective group.

• good. The total number of 'Goods' or 'Events' in respective group.

• bad. The total number of 'Bads' or 'Non-Events' in respective group.

• pert. The Percentage of observations accounted for by respective group.

• pertg. The Percentage of 'Goods' or 'Events' accounted for by respective group.

• pertb. The Percentage of 'Bads' or 'Non-Events' accounted for by respective group.

• br. The Percentage of 'Bads' or 'Non-Events' in respective group.

• woe. The computed weights of evidence(WOE) for respective group. The WOE values can be used in place of the actual group itself, thereby producing a 'continuous' alternative.

• iv. The information value contributed by each group in the X. The sum of IVs is the total information value of the categorical X variable.

Author(s)

Nguyen Ngoc Binh [e-mail]

Examples

data('hmeq')
hmeq1 <- hmeq[,c("bad", "reason", "job")]
woe_table <- woe_cal_table(df = hmeq1, y = "bad")

nguyenngocbinh/scorecard.sea documentation built on Jan. 4, 2020, 11:43 a.m.