LOOCV: leave-one-out cross-validation

In cma2015/RAP: Rank aggregation-based data fusion for gene prioritization

Description

Leave-one-out cross-validation algorithm is performed to train and test the integrative random-forest gene prioritization algorithm RafSee.

Usage

  LOOCV(featureMat, positives, negatives, cpus = 1, predictSample = NULL)

Arguments

featureMat	A numeric matrix of features where rows represent genes, cols represent features
positives	A character vector of positive samples
negatives	A character vector of negative samples
cpus	an integer number specifying the number of cpus to be used for parallel computing, the default is 1
predictSample	A vector of testing samples, if it is NULL, all genes excluding positive smaples were used

Value

Predictive score for each leave-one-out cross-validation

Author(s)

Jingjing Zhai, Chuang Ma

Examples

  ## Not run: 
    positives <- c("AT1G01060", "AT1G09530", "AT1G09570", "AT1G12610")
    loocvRes <- LOOCV(featureMat = featureMat, positives = positives, 
                      negatives  = negatives, cpus = 1)
    ## featureMat can be calculated by function FeatureExtract
    ## negatives can be calculated by function selectNegSamples
  
## End(Not run)

cma2015/RAP documentation built on Aug. 18, 2019, 8:02 p.m.