find_k {jackstraw}

Find a number of clusters or principal components

Description

There are a wide range of algorithms and visual techniques to identify a number of clusters or principal components embeded in the observed data.

Usage

find_k()

Details

It is critical to explore the eigenvalues, cluster stability, and visualization. See R packages bootcluster, EMCluster, and nFactors.

Please see the R package SC3, which provides estkTW() function to find the number of significant eigenvalues according to the Tracy-Widom test.

ADPclust package includes adpclust() function that runs the algorithm on a range of K values. It helps you to identify the most suitable number of clusters.

This package also provides an alternative methods in permutationPA. Through a resampling-based Parallel Analysis, it finds a number of significant components.