Skip to content

Why TARNet loss only mse? #16

@cancermqiao

Description

@cancermqiao

In TARNet builder, there are two outs in model_outputs: propensity score and potential outcomes. Why the model compile only use mse loss? Whether the cross-entropy loss should be used for the propensity score?
image

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions