Detecting eigengap.

Searching for the largest gap between conjuncted eigenvalues which are sorted in an ascending order, the most important eigenvalues and eigenvectors are the ones before the detected eigengap.

Usage

eigengap(e, v)

Arguments

e

Eigenvalues of type 'vector'

v

Eigenvector matirx of type 'matrix/array' whose columns are single eigenvectors.

Value

The selected most important eigenvectors for downstream clustering, of type 'matrix/array'.