stability: Refit models for stability selection

In amcdavid/HurdleNormal: 'Estimation and sampling for multi-dimensional Hurdle models on a Normal density with applications to single-cell co-expression'

Refit models for stability selection

Description

The function method (which needs to follow the API of autoGLM) is called on subsampled data. Exceptions are caught and output saved to disk since this can be quite computationally expensive.

Usage

stability(
  obs,
  fixed,
  stabIndex,
  step = seq_along(stabIndex),
  method,
  stabilityCheckpointDir = NULL,
  checkpointDir = NULL,
  ...
)

Arguments

obs	a matrix of observations from which rows will be sampled
fixed	a matrix of covariates
stabIndex	output from setupStabilityIndex
step	indices of components to run from stabIndex. Defaults to all.
method	what method, eg, fitHurdle or autoGLM
stabilityCheckpointDir	path to save output from each stability iteration
checkpointDir	path to save intermediate output within each stability iteration
...	arguments passed to method

Value

list of output from method, eg, adjacency matrices.

amcdavid/HurdleNormal documentation built on May 14, 2022, 11:12 p.m.