train_rf: Generate six RF models from two classes

In g1o/GeneEssentiality: Gene essentiality predictor

Description

This returns a list with one random forest model

Usage

train_rf(
  features = Complete_set,
  CPU = 2,
  trees = 1000,
  CV = 10,
  nrepeats = 3,
  seeds = seed,
  saveprediction = "final"
)

Arguments

CPU	Number of threads to use when training the model
trees	Number of trees for the random forest (rf) model
CV	Number of sets for the cross-validation
nrepeats
data	A feature data frame including the label of two Classes

Details

The followind had no significant improvement and took too much time. In the future it may be activated with another parameter. It will return a list with 3 normal rf models and 3 rf models made with randomized labels. The models 1 and 2 are trained without any subsampling. #' The models 3 and 4 are trained using upsampling during the cross-validation (Subsampling During Resampling). The models 5 and 6 are trained using downsampling during the cross-validation (Subsampling During Resampling).

Value

A list of models

g1o/GeneEssentiality documentation built on Jan. 3, 2022, 1:21 a.m.