Most Artificial Intelligence models require the information in the records to be used to be fully informed. These models require a policy for handling missing information. However, traditional policies have tried to fill in the missing information with known information. This approach is correct when the missing information is random and wrong when it is not. In this article we start with a review of the main policies employed, and analyse the consequences of elimination or imputation policies and argue that these policies are sometimes unwise.
Except where otherwise noted, this item's license is described as Attribution-NonCommercial-NoDerivatives 4.0 Internacional