use_glmnet.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/use_glmnet.R
\name{use_glmnet}
\alias{use_glmnet}
\title{Use glmnet}
\usage{
use_glmnet(
  data = NULL,
  assay_classification = NULL,
  classification = NULL,
  model_matrix = NULL,
  col_names = NULL,
  gene_names = NULL,
  use_glm = T,
  use_cv = T,
  fast_cv = F,
  stratify_class = F,
  type_measure = "class",
  type.multinomial = "grouped",
  alpha = 0.5,
  verbose = T,
  parallel = T,
  no_cluster = NULL,
  seed = NULL,
  repeats = 1,
  k_outer_loop = 10,
  k_inner_loop = 5,
  standardize = F,
  return_both = T,
  output_genewise = T,
  ...
)
}
\arguments{
\item{data}{A data frame or SummarizedExperiment.}

\item{assay_classification}{Integer - number of assay that contains the classification made by mclust}

\item{classification}{Matrix - single named vector or matrix of the classification of the samples}

\item{model_matrix}{Matrix - a model matrix as for example generated by /link [make_model_matrix]}

\item{col_names}{Character - set of column names to be used as predictors in regression}

\item{gene_names}{Character - vector of gene names to apply glmnet to}

\item{use_glm}{Do a signle glmnet fit?}

\item{use_cv}{Logical - should nested cross validation be used to estimate error?}

\item{fast_cv}{Logical - faster calc because entries with zero coefficients are skipped.}

\item{verbose}{Logical - Show Progress of function, not in parallel}

\item{seed}{Integer - Number for set.seed}

\item{output_genewise}{Logical - Genewise results as list, otherwise a list with glm and cv results with all genes}

\item{...}{Parameters for glmnet}

\item{return_plot}{Logical - If set to true, the output is a list of the Regression  results and a correlation plot with AUC plot.}

\item{exclude}{character vector - factors/columns to be excluded. You dont want to include factors which are more like observatory variables as opposed to possible explanatory variables in the analysis.}
}
\description{
This function uses glmnet to do variable selection within a classification of a response variable.
}
\details{
Best use for summarized experiments. Output genewise T is better.
}