getFeatureCorrelations: Compute correlations between a data matrix and a signal...
In tkonopka/Rthresholdout: Re-usable holdout implementation in R

Evaluates correlations between features in a data matrix and a signal vector. When only one data and signal object is provided, the output is a vector of straightforward correlations. When a background dataset and background signal vector are also provided, the function treats the primary data as a reusable holdout: it will output correlations from either the holdout data or the background set.

1
2
3

getFeatureCorrelations(dat, signal, min.cor = 1/sqrt(length(signal)),
  dat.bg = NULL, signal.bg = NULL, tolerance.factor = 1,
  threshold.factor = 4)

`dat`	- a data matrix with S samples in columns and F features rows
`signal`	- a numeric vector. The function will compute correlations between rows in the data matrix and this signal vector.
`min.cor`	- minimal expected correlation (function will output zero if the actuall correlation is below threshold)
`dat.bg`	- background dataset
`signal.bg`	- background signal
`tolerance.factor`	- one of the penalties used in the reusable holdout proposal
`threshold.factor`	- one of the penalties used in the reusable holdout proposal.