Random_spongEffects: build random classifiers

In biomedbigdata/SPONGE: Sparse Partial Correlations On Gene Expression

build random classifiers

Description

build random classifiers

Usage

Random_spongEffects(
  sponge_modules,
  gene_expr,
  min.size = 10,
  bin.size = 100,
  max.size = 200,
  min.expression = 10,
  replace = F,
  method = "OE",
  cores = 1
)

Arguments

sponge_modules	result of define_modules()
gene_expr	Input expression matri
min.size	minimum module size (default: 10)
bin.size	bin size (default: 100)
max.size	maximum module size (default: 200)
replace	Possibility of keeping or removing (default) central genes in the modules (default: F)
method	Enrichment to be used (Overall Enrichment: OE or Gene Set Variation Analysis: GSVA) (default: OE)
cores	number of cores to be used to calculate entichment scores with gsva or ssgsea methods. Default 1
train_gene_expr	expression data of train dataset, genenames must be in rownames
test_gene_expr	expression data of test dataset, genenames must be in rownames
train_meta_data	meta data of train dataset
test_meta_data	meta data of test dataset
train_meta_data_type	TCGA or METABRIC
test_meta_data_type	TCGA or METABRIC
metric	metric (Exact_match, Accuracy) (default: Exact_match)
tunegrid_c	defines the grid for the hyperparameter optimization during cross validation (caret package) (default: 1:100)
n.folds	number of folds to be calculated
repetitions	number of k-fold cv iterations (default: 3)
min.expr	minimum expression (default: 10)

Value

randomized prediction model Define random modules

A list with randomly defined modules and related enrichment scores

biomedbigdata/SPONGE documentation built on Feb. 6, 2023, 10:19 p.m.