Train the Outcome Model (Y)

Jointly trains model_y (causal margin) and model_eta (association remainder). Uses the Z-permutation trick to ensure eta becomes marginally N(0,I).

Usage

train_y(model, ...)

# S3 method for class 'frengression'
train_y(
  model,
  x,
  z,
  y,
  num_iters = 100,
  lr = 0.001,
  print_every = 10,
  tol = 0.01,
  silent = FALSE,
  ...
)

# S3 method for class 'frengression_seq'
train_y(
  model,
  s,
  x,
  z,
  y,
  num_iters = 100,
  lr = 1e-04,
  print_every = 10,
  silent = FALSE,
  ...
)

# S3 method for class 'frengression_surv'
train_y(
  model,
  s,
  x,
  z,
  y,
  num_iters = 100,
  lr = 0.001,
  print_every = 10,
  reg_lambda = 0,
  silent = FALSE,
  ...
)

Arguments

model

A frengression, frengression_seq, or frengression_surv object.

...

Additional arguments passed to methods.

x

Treatment matrix or tensor.

z

Confounder matrix or tensor.

y

Outcome matrix or tensor (events).

num_iters

Number of training iterations (default: 100).

lr

Learning rate (default: 1e-3).

print_every

Number of iterations between progress prints (default: 10).

tol

Early stopping tolerance (default: 0.01).

silent

Suppress output (default: FALSE).

s

Static baseline variables (matrix or tensor).

reg_lambda

Regularization weight for marginal loss (default: 0).

Value

The model with updated weights and training loss history.