sgd_given_c: Runs a stochastic gradient descent to minimize SURE for...
In dcgerard/hose: Higher-order spectral estimators

Description Usage Arguments Author(s) References

This function is in beta. Not sure if it works properly or is even worthwhile considering that soft_coord works pretty well. Also, we have to call get_c anyway, so not sure if this function actually reduces the computational complexity.

sgd_given_c(c_obj, sgd_lambda, sgd_lambda_c, c_init = 1,
  lambda_init = NULL, sgd_c = 1/2 + 0.001, itermax = 10000,
  tau2 = 1, print_current = TRUE, every_iter = 1000, alpha = 0.2,
  calc_final = TRUE)

`c_obj`	= Output from `get_c`.
`sgd_lambda`	Lambda value from sgd for lambda from hose.
`sgd_lambda_c`	Lambda value from sgd for c from hose.
`c_init`	Initialization for c from hose.
`lambda_init`	Initialization for lambda from hose.
`sgd_c`	c value from sgd.
`itermax`	Maximum number of iterations for sgd.
`tau2`	Known variance.
`print_current`	Should we print the results at each iteration?
`every_iter`	If `print_current = TRUE`, then every `every_iter` iteration will be printed.
`alpha`	Burnin for final estimates.
`calc_final`	Should we calculate the final sure and estimates?