rdcPart: Randomized dependence coefficient partial calculation
In kernDeepStackNet: Kernel Deep Stacking Networks

Description Usage Arguments Details Value Note Author(s) References See Also

Calculates the randomized dependence coefficient based on interim results. It is a generalized dependence measure based on maximum correlation of random non-linear projections.

1	rdcPart(subsetX, xTrans, yTrans, s=1/6, f=sin, randX)

`subsetX`	Subset of the covariate matrix as indices (integer vector).
`xTrans`	Transformed matrix to the [0, 1] scale (numeric matrix).
`yTrans`	Random, non-linear projection of the response (numeric vector).
`s`	Variance of the random weights. Default is 1/6.
`f`	Non-linear transformation function. Default is `sin`.
`randX`	Random weights (numeric vector).

This function allows for more efficient calculation than the complete calculation by excluding repetitive calculations.

Value of randomized dependence coefficient (numeric scalar).

This function is a help function within variable selection. It is given for experts for model customization. It is recommended to use instead function rdcVarOrder.

Thomas Welchowski welchow@imbie.meb.uni-bonn.de

David Lopez-Paz et. al, (2013), The randomized dependence coefficient, Proceedings of Advances in Neural Information Processing Systems 26 (NIPS)

rdcVarOrder, cancorRed

kernDeepStackNet documentation built on May 2, 2019, 8:16 a.m.