preselect: Create a list of variables using various variable selection...

Description Usage Arguments

View source: R/preselect.R

Description

preselect

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
preselect(
  input,
  operations,
  replicates = NULL,
  batch_id = NULL,
  subset = NULL,
  cores = NULL,
  output_dir = NULL,
  workspace_dir = "."
)

Arguments

input

Test data on which to perform variable selection operations. Must be CSV, rds, or feather.

operations

List of operations to perform, as strings. Supported operations: "correlation_threshold", "variance_threshold", "replicate_correlation". See cytominer::select.

replicates

Number of replicates to select per plate map. Required for the operation "replicate_correlation". See cytominer::preselect. default: NULL.

batch_id

Batch ID. Used for generating output_dir if the latter is not specified. default: NULL.

subset

Query to create the training data by subsetting. Regex must select CSVs only. Default selects everything. default: NULL.

cores

Optional integer specifying number of CPU cores used for parallel computing using doParallel. default: NULL.

output_dir

Output directory for preselected feature names. By default, writes to workspace_dir/parameters/batch_id/variable_selection/. default: NULL.

workspace_dir

Root directory containing backend and metadata subdirectories. Can be relative or absolute. default: ".".


cytomining/cytotools documentation built on Nov. 19, 2020, 10:23 p.m.