find_sigmas: Calculate the average dimensionality for m different gaussian...

In theislab/destiny: Creates diffusion maps

Calculate the average dimensionality for m different gaussian kernel widths (\sigma).

Description

The sigma with the maximum value in average dimensionality is close to the ideal one. Increasing step number gets this nearer to the ideal one.

Usage

find_sigmas(
  data,
  step_size = 0.1,
  steps = 10L,
  start = NULL,
  sample_rows = 500L,
  early_exit = FALSE,
  ...,
  censor_val = NULL,
  censor_range = NULL,
  missing_range = NULL,
  vars = NULL,
  verbose = TRUE
)

Arguments

data	Data set with n observations. Can be a data.frame, matrix, ExpressionSet or SingleCellExperiment.
step_size	Size of log-sigma steps
steps	Number of steps/calculations
start	Initial value to search from. (Optional. default: \log_{10}(min(dist(data))))
sample_rows	Number of random rows to use for sigma estimation or vector of row indices/names to use. In the first case, only used if actually smaller than the number of available rows (Optional. default: 500)
early_exit	logical. If TRUE, return if the first local maximum is found, else keep running
...	Unused. All parameters to the right of the ... have to be specified by name (e.g. find_sigmas(data, verbose = FALSE))
censor_val	Value regarded as uncertain. Either a single value or one for every dimension
censor_range	Uncertainity range for censoring. A length-2-vector of certainty range start and end. TODO: also allow 2\times G matrix
missing_range	Whole data range for missing value model. Has to be specified if NAs are in the data
vars	Variables (columns) of the data to use. Specifying TRUE will select all columns (default: All floating point value columns)
verbose	logical. If TRUE, show a progress bar and plot the output

Value

Object of class Sigmas

See Also

Sigmas, the class returned by this; DiffusionMap, the class this is used for

Examples

data(guo)
sigs <- find_sigmas(guo, verbose = TRUE)
DiffusionMap(guo, sigs)

theislab/destiny documentation built on Oct. 30, 2024, 1:39 a.m.