bpr_optimize {BPRMeth}R Documentation

(DEPRECATED) Optimize BPR negative log likelihood function

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:

Author(s)

C.A.Kapourani C.A.Kapourani@ed.ac.uk

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)


[Package BPRMeth version 1.6.0 Index]