find_closer_points {archetypal}

Function which finds the data points that are closer to the archetypes during all iterations of the algorithm PCHA

Description

This function runs the PCHA algorithm and finds the data points that are in the local neighborhood of each archetype. The size of the neighborhood is user defined (npoints). This allows us to study the properties of the solution or manually choose an initial approximation to search for a better fit.

Usage

find_closer_points(df, kappas, usedata = FALSE, npoints = 2, 
                     nworkers = NULL, rseed = NULL, 
                     verbose = FALSE, doparallel = FALSE, ...)

Arguments

df

The data frame with dimensions n x d

kappas

The number of archetypes

usedata

If it is TRUE, then entire data frame will be used, if doparallel = TRUE

npoints

The number of closer points to be estimated

nworkers

The number of logical processors that will be used, if doparallel = TRUE

rseed

The random seed that will be used for random generator. Useful for reproducible results.

verbose

If it is set to TRUE, then details will be printed, except from archetypal

doparallel

If it is set to TRUE, then parallel processing will be performed

...

Other arguments to be passed to archetypal except internally used save_history = TRUE and verbose = FALSE. This is essential for using optimal parameters found by find_pcha_optimal_parameters

Value

A list with members:

1. rows_history, a list with npoints rows used that are closer to each archetype for each iteration done by algorithm

2. iter_terminal, iteration after which rows closer to archetypes do not change any more

3. rows_closer, the rows closer to archetypes by means of Euclidean distance and are fixed after iter_terminal iteration

4. rows_closer_matrix, a matrix with npoints rows which are closer to each archetype

5. solution_used, the AA output that has been used. Some times useful, especially for big data.

See Also

check_Bmatrix, study_AAconvergence

Examples

{
# Load data "wd2"
data("wd2")
yy = find_closer_points(df = wd2, kappas = 3, npoints = 2, nworkers = 2)
yy$rows_history
yy$iter_terminal
yy$rows_closer
yy$rows_closer_matrix
yy$solution_used$BY

}