sig.explore: Estimate optimal sigma parameter.

In hms-dbmi/crestree: Analysis of branching branching trajectories from single-cell data

Description

Using cross-validation criteria to select sigma parameter.

Usage

sig.explore(X, W = NA, M = as.integer(ncol(X)/2), n.sample = 1,
  sig.lims = seq(0.01, 0.2, 0.03), metrics = "cosine", p.power = 1,
  plot = TRUE, err.cut = 0.5, n.steps = 20, n.cores = 1)

Arguments

X	gene (rows) vs cell (columns) expression matrix
M	number of principal points in pptree modeling
n.sample	number of sampled trees per each sigma
sig.lims	a vector of sigma for which cross-validation estimated
metrics	similarity measure. "cosine" or "euclidean"

Value

optimal sigma parameter

hms-dbmi/crestree documentation built on July 11, 2019, 8:04 p.m.