My target variable is continuous in the [0, 1] interval, what the best model for this I can build in pytorch? I am wondering if there’s some form of logistic regression (with appropriate loss function) that works for this?
To keep the outcome within [0,1]
, you can formulate any model with the final layer and the objective functions as follows:
- Final layer:
Sigmoid
, Objective function:BCELoss
, - Final layer:
Linear
, Objective function:BCEWithLogitsLoss