https://packaging.python.org/en/latest/

Steps:

1. Obtain the center in 5D
2. Select the 25th (?) closest stars for each dimension
3. Use the distance to that star to normalize each dimension separately
4. Obtain the 5D euclidean distance to the center for the normalized dimensions
5. Order all stars according to this distance
6. Calculate the density from the 25th to the 5000th star in steps of 10 (?)
7. Select the optimal number according to some criteria (drop in density?)

Test using Voronoi volumes for estimating the number of stars beyond the volume limit of a random/uniform distribution of stars

Detecting Significant Multidimensional Spatial Clusters

The Application of the Multi-Dimensional Clustering Algorithm in the Field of Abell 383

Currently the code select the N_memb stars closest to the center with no photometric input. The G magnitude can be used to estimate the number of stars in 2, 3, or 4 ranges of its distribution, and then select the proper number of stars in each region.

For example, if the estimated N_memb=100, and there are 15 stars in the first percentile of the G mag (0%-33%, brightest stars), 35 in the second (33%-66%) and 50 in the third (66%-100%, dimmest stars) then select the proper amount in each region as the closest to the centers.