sigma_finder {Spectrum}

sigma_finder: heuristic to find sigma for the Ng kernel

Description

This is a heuristic to find the sigma for the kernel from the Ng spectral clustering algorithm. It returns a global sigma. It uses the mean K nearest neighbour distances of all samples to determine sigma.

Usage

sigma_finder(mat, NN = 3)

Arguments

mat

Data frame or matrix: with points as columns, features as rows

NN

Numerical value: the number of nearest neighbours to use (default=3)

Value

A global sigma

Examples

sig <- sigma_finder(blobs)