Variables transformation for Scorecard development

[ggfggrr]

I am currently building a scorecard for a risk model. While trying to improve the model, I came across an option to transforming the variables (for eg. standardization) before grouping. Wish to receive an advise whether this can be done and considered as an option? If yes, what precautions should be.

Thanks

Kind regards,
Mari

[Ksharp]

One way to improve your model is making your variable look like normal distribution, so try to use function LOG() .