check_Bmatrix function

Function which checks B matrix of Archetypal Analysis Y ~ A B Y in order to find the used rows for creating each archetype and the relevant used weights.

check_Bmatrix(B, chvertices = NULL, verbose = TRUE)

Arguments

• B: The $kappas \times n$ matrix such that Y ~ ABY or Frobenius norm ||Y-ABY|| is minimum
• chvertices: The vector of rows which represent the Convex Hull of data frame
• verbose: If set to TRUE, then results are printed out.

Returns

A list with members:

1. used_rows, a list with used rows for creating each archetype
2. used_weights, a list with the relevant weights that have been used
3. leading_rows, the rows for each archetype with greatest weight
4. leading_weights, the weights of leading rows
5. used_on_convexhull, the portion of used rows which lie on Convex Hull (if given)

Examples

{
        # Load data "wd2"
        data("wd2")
        df = wd2
        # Run AA:
        aa = archetypal(df = df, kappas = 3, verbose = FALSE)
        # Check B matrix:
        B = aa$B
        yy = check_Bmatrix(B, verbose = TRUE)
        yy$used_rows
        yy$used_weights
        yy$leading_rows
        yy$leading_weights
        # Check if used rows lie on ConvexHull
        ch = chull(df)
        yy = check_Bmatrix(B, chvertices = ch, verbose = FALSE)
        yy$used_on_convexhull
        #

}

See Also

archetypal, check_Bmatrix, find_closer_points

& study_AAconvergence