Imputation

Imputation is one of the core preprocessing techniques for handling missing data. It can range from simple mean/median filling to advanced model-based imputation. However, every imputation method introduces assumptions into the data, which is why it must be applied carefully. A poor imputation strategy may distort distributions, alter relationships, and create artificial regularity. For that reason, the choice of imputation method should depend on why the data is missing, the type of variable, and the modeling objective.