Which of the following most closely describes embedded methods of feature selection?

involve proposing predictor candidates based on the analyst's domain knowledge

examine the correlation of each variable with the outcome variable

reduce the variable selection by examining how useful each one is in the model's performance

automatically include the variable selection as part of the model's optimization algorithm