train_model: Model training

In elimereu/matchSCore2: What the Package Does TODOELI

Description

This function train a model from the reference dataset

Usage

train_model(
  scale.data,
  clus,
  gene_cl.ref,
  prop = NULL,
  p.threshold = NULL,
  verbose = TRUE,
  ...
)

Arguments

scale.data	A scaled matrix of gene expressions like in the scale.data of the Seurat object. Rows are genes and columns are cells from the reference dataset.
clus	A factor with identities from the reference dataset.
gene_cl.ref	A list of cluster-specific markers. Each element of the list contains markers of a cell type. Usually only top100 ranked markers are used.
prop	The proportion of the reference data used to train the model. Default=0.5.
p.threshold	Probability threshold to consider a cell classified. Default=0.65.
verbose	Logical, controls the displaying of additional messages while running the function. Defaults to TRUE.
...	TODOELI

Value

A multinomial fitted model, as in the nnet package.

Examples

# TODO

elimereu/matchSCore2 documentation built on April 9, 2020, 5:41 p.m.