bpr_optimize: (DEPRECATED) Optimize BPR negative log likelihood function

In andreaskapou/BPRMeth: Model higher-order methylation profiles

Description

(DEPRECATED) The function bpr_optimize minimizes the negative log likelihood of the BPR function. Since it cannot be evaluated analytically, an optimization procedure is used. The optim packages is used for performing optimization.

Usage

bpr_optim(x, ...)

## S3 method for class 'list'
bpr_optim(
  x,
  w = NULL,
  basis = NULL,
  lambda = 1/2,
  opt_method = "CG",
  opt_itnmax = 100,
  is_parallel = TRUE,
  no_cores = NULL,
  ...
)

## S3 method for class 'matrix'
bpr_optim(
  x,
  w = NULL,
  basis = NULL,
  lambda = 1/2,
  opt_method = "CG",
  opt_itnmax = 100,
  ...
)

Arguments

x	The input object, either a matrix or a list.
...	Additional parameters.
w	A vector of parameters (i.e. coefficients of the basis functions)
basis	A 'basis' object. E.g. see create_rbf_object.
lambda	The complexity penalty coefficient for ridge regression.
opt_method	The optimization method to be used. See optim for possible methods. Default is "CG".
opt_itnmax	Optional argument giving the maximum number of iterations for the corresponding method. See optim for details.
is_parallel	Logical, indicating if code should be run in parallel.
no_cores	Number of cores to be used, default is max_no_cores - 2.

Value

Depending on the input object x:

• If x is a list: An object containing the following elements:

  • W_opt: An Nx(M+1) matrix with the optimized parameter values. Each row of the matrix corresponds to each element of the list x. The columns are of the same length as the parameter vector w (i.e. number of basis functions).

  • Mus: An N x M matrix with the RBF centers if basis object is create_rbf_object, otherwise NULL.

  • basis: The basis object.

  • w: The initial values of the parameters w.

• If x is a matrix: An object containing the following elements:

  • w_opt: Optimized values for the coefficient vector w. The length of the result is the same as the length of the vector w.

  • basis: The basis object.

• If calling bpr_optim_fast just the optimal weight matrix W_opt.

Author(s)

C.A.Kapourani [e-mail]

See Also

create_basis, eval_functions

Examples

# Example of optimizing parameters for synthetic data using default values
data <- meth_data
out_opt <- bpr_optim(x = data, is_parallel = FALSE, opt_itnmax = 3)

#-------------------------------------

# Example of optimizing parameters for synthetic data using 3 RBFs
ex_data <- meth_data
basis <- create_rbf_object(M=3)
out_opt <- bpr_optim(x = ex_data, is_parallel = FALSE, basis = basis,
                     opt_itnmax = 3)

#-------------------------------------

# Example of of specific promoter region using 2 RBFs
basis <- create_rbf_object(M=2)
w <- c(0.1, 0.1, 0.1)
data <- meth_data[[1]]
out_opt <- bpr_optim(x = data, w = w, basis = basis, fit_feature = "NLL",
                     opt_itnmax = 3)

andreaskapou/BPRMeth documentation built on June 11, 2020, 10:49 p.m.