Credit Risk Modelling
Hi Team,Trust your are doing well !
In PD modelling, why do we need to convert continuous variables to categories ? What if we remain continuous variable as it is and encode categorical variables using one hot encoding/other encoding techniques. If this can be possible, what is the impact of this on Population Stability Index ?
1 answers ( 0 marked as helpful)
Hi Phanindra,
Great question!
Normally in data science, we would not want to do that. As you are probably assuming, the accuracy would be much higher if we could simply use continuous variables.
However, due to regulation (namely from Basel II) we must include only categorical variables in a PD model so that it is easy to interpret.
Best,
Iliya