kmeansDS: Parallel k-means iteration

In isglobal-brge/dsML: DataSHIELD server site machine learning functions.

Parallel k-means iteration

Description

Performs an iteration of a k-means parallel algorithm (what in a multi-thread machine would be run on each thread). The client acts as the master and the servers as the slaves if thinking like a regular parallel implementation.

Usage

kmeansDS(x, ...)

Arguments

x	data frame Train dataset for the k-means
...	numeric Parameters corresponding to the data frame from the server that contains the centroids (updated on each iteration on the client)

Value

list with:
-counts numeric vector with the counts per cluster
-centers data frame New centroids calculated
-assignations numeric vector ordered cluster assignations, to be used by the client to assign them with the kmeans.assign_result function to a table on the servers to be later used

isglobal-brge/dsML documentation built on March 14, 2023, 1:58 p.m.