successiveOmitClusterValidation: Cluster Validation by Successive Omission

In nguforche/UnsupRF: Unsupervised Random Forest Clustering

Description

Validate a clustering scheme by succerssively omitting the most compact/connected cluster member and re-cluster the remaining data points, and continue the process all over. The hypothesis is that, if the data can be clustered into k optimal cluster, then after dropping the most compact/connected cluster member, the remaining data points can best be clustered into k-1 clusters and so on, untill k = 2. By trying different k's, the successive omission algorithm may identify the best k clusters for the data.

Usage

successiveOmitClusterValidation(data, k = 2, vars,
  cluster.method = "UnsupRF", dist.method = "euclidean",
  control = NULL, ensemble = TRUE, sep.measure = "SW")

Details

This function is experimental and not meant to be used directly by the user Work in progress to validate the procedure

nguforche/UnsupRF documentation built on May 5, 2019, 4:51 p.m.