Autobinning response highest count gets to be good?

2 次查看(过去 30 天)
Hi I read in the documentation of the credit risk toolkit that when using binning the response value with the most occurencies get's to be the 'Good' one.
Is there any way to change that?

回答(1 个)

Michael
Michael 2024-8-27
Yes, this can be changed. As you noted, the default GoodLabel is the response value with the highest count. The optional parameter GoodLabel can be used to explicitly set which value is the "Good" one.
As an example, try the following code:
load("CreditCardData.mat")
% Check that 0 is the most common
responseCounts = groupsummary(data, "status");
disp(responseCounts)
% Check that the default is to set the GoodLabel to 0
scDefault = creditscorecard(data);
disp(scDefault.GoodLabel)
% Check that you can change the GoodLabel to 1
scNonDefault = creditscorecard(data, "GoodLabel", 1);
disp(scNonDefault.GoodLabel)

类别

Help CenterFile Exchange 中查找有关 Probability Distributions and Hypothesis Tests 的更多信息

产品


版本

R2018b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by