Apart from heuristically treating an encoder like a hyperparameter, how can one decide which encoder to use for continuous features? Rule out encoders that do/ don't accept negative values. Can we look at stdev and skewness to determine which encoder would provide the best transformation?

Factors in choosing a continuous encoder?

Data Science Asked on March 3, 2021

Apart from heuristically treating an encoder like a hyperparameter, how can one decide which encoder to use for continuous features?

Rule out encoders that do/ don’t accept negative values.
Can we look at stdev and skewness to determine which encoder would provide the best transformation?

Get help from others!