df_ncut: Gradient of the Normalised Cut Across a Hyperplane

In DavidHofmeyr/PPCI: Projection Pursuit for Cluster Identification

Description

Finds the gradient of the normalised cut across the best hyperplane orthogonal to a given projection vector. Used to obtain minimum normalised cut hyperplanes using gradient based optimisation.

Usage

df_ncut(v, X, P)

Arguments

v	a numeric vector of length ncol(X)
X	a numeric matrix (num_data x num_dimensions) to be projected on v
P	a list of parameters including (at least) $s (positive numeric scaling parameter), $nmin (positive integer minimum cluster size).

Value

the (vector) gradient of the normalised cut across the best hyperplane orthogonal to v.

DavidHofmeyr/PPCI documentation built on March 9, 2020, 5:05 p.m.