subClass_train: Sub-Class Training (new and improved version) with merged...

In pcahan1/cancerCellNet: CellNet, for cancer

Sub-Class Training (new and improved version) with merged rand category

Description

Tranining sub-class classifier

Usage

subClass_train(
  cnProc_broad,
  stTrain,
  expTrain,
  colName_broadCat,
  colName_subClass,
  name_broadCat,
  colName_samp = "row.names",
  nTopGenes = 20,
  nTopGenePairs = 50,
  nRand = 40,
  nTrees = 1000,
  stratify = FALSE,
  sampsize = 40,
  weightedDown_total = 5e+05,
  weightedDown_dThresh = 0.25,
  transprop_xFact = 1e+05,
  weight_broadClass = 1,
  quickPairs = FALSE,
  coreProportion = 0
)

Arguments

cnProc_broad	the broad cnProc from broadClass_train
stTrain	a dataframe that matches the samples with broad category and sub-class
expTrain	the expression matrix
colName_broadCat	the name of the column in sample table that contains broad categories
colName_subClass	the name of the column in sample table that contains sub class
name_broadCat	the name of the broad class in which the subclasses are
colName_samp	the name of the column that contains sample names
nTopGenes	the number of classification genes per category
nTopGenePairs	the number of top gene pairs per category
nRand	number of random profiles generate for training
nTrees	number of trees for random forest classifier
weightedDown_dThresh	the threshold at which anything lower than that is 0 for weighted_down function
transprop_xFact	scaling factor for transprop
weight_broadClass	the weight on the result of the broad classification as features for subclassifier
coreProportion	the proportion of logical cores for finding classification genes and top scoring gene pairs. If you want to disable parallel processing, then enter 0
weightDown_total	numeric post transformation sum of read counts for weighted_down function

Value

a list containing normalized expression data, classification gene list, cnProc

pcahan1/cancerCellNet documentation built on July 16, 2022, 12:12 a.m.