rRedundantPredictors: Remove redundant predictors

In carmen-maria-hernandez/coexnetwork: Development of supervidsed co-expression networks

Description

Reduces the number of predictors by eliminating those with near-zero variance and from pairs that have high correlation (>= 90%) eliminate one of the two predictors.

Usage

rRedundantPredictors(data)

Arguments

data	numerical matrix with predictors as rows and samples as columns

Value

numerical matrix

Examples

nycfl = nycflights13::flights
# We won´t use those non numerical predictors
nycfl = na.omit(nycfl[,-c(1,10,12,13,14,19)])
ncol(nycfl)
nycfl2 <- t(rRedundantPredictors(t(nycfl)))
ncol(nycfl2)

carmen-maria-hernandez/coexnetwork documentation built on Dec. 19, 2021, 1:53 p.m.