obsEval: Observation-wise Clustering Robustness Evaluation

Description Usage Arguments Value Author(s) Examples

Description

A sample observation-wise clustering robustness evaluation framework (described in "Examples" section, used as default in iterClust framework). Customized frameworks can be defined following rules specified in "Usage", "Arguments" and "Value" sections.

Usage

1
obsEval(dset, clust, iteration)

Arguments

dset

(numeric matrix) features in rows and observations in columns

clust

optimal return value of coreClust

iteration

(positive integer) specifies current iteration

Value

a numeric vector, specifies the clustering robustness (higher value means more robust) of each observation under the optimal clustering scheme

Author(s)

DING, HONGXU (hd2326@columbia.edu)

Examples

1
2
3
4
obsEval <- function(dset, clust, iteration){
    dist <- as.dist(1 - cor(dset))
    obsEval <- vector("numeric", length(clust))
    return(silhouette(clust, dist)[, "sil_width"])}

hd2326/iterClust documentation built on May 31, 2019, 3:54 a.m.