A New WOE and Scorecard Building R Package
Presentation Type
Event
Abstract
In almost every data science project that involves modeling, cleaning and transforming the data is the most time-consuming part. One method of data transformation is called weights of evidence (WOE) binning. WOE binning is commonly used with credit scoring because most of the data is highly skewed and contains numerous missing values. Despite WOE binning being a common practice in the credit scoring industry, there are relatively few R packages that perform WOE binning and also build a scorecard. Due to the lack of packages in R capable of these methods, we believed it would be beneficial to the R community to build an R package that performed WOE binning and scorecard creation. The R package created has built-in functions to allow the user to complete all parts of the credit scoring process. These functions include: custom binning, WOE transformation, variable clustering, scorecard creation, and WOE visualization. By creating this R package, users will have the ability to create a scorecard using WOE transformation for their data science projects.
Start Date
2-12-2018 12:00 PM
A New WOE and Scorecard Building R Package
In almost every data science project that involves modeling, cleaning and transforming the data is the most time-consuming part. One method of data transformation is called weights of evidence (WOE) binning. WOE binning is commonly used with credit scoring because most of the data is highly skewed and contains numerous missing values. Despite WOE binning being a common practice in the credit scoring industry, there are relatively few R packages that perform WOE binning and also build a scorecard. Due to the lack of packages in R capable of these methods, we believed it would be beneficial to the R community to build an R package that performed WOE binning and scorecard creation. The R package created has built-in functions to allow the user to complete all parts of the credit scoring process. These functions include: custom binning, WOE transformation, variable clustering, scorecard creation, and WOE visualization. By creating this R package, users will have the ability to create a scorecard using WOE transformation for their data science projects.