NAME TO BE DETERMINED

This project began on June 2022.

TODO

1. Process large files

Cantat-Gaudin data

• Cantat-Gaudin & Anders (2020):

1481 clusters (435833 stars with P>0.01); G_max=18

Average number of members

For members identified as those with P>0.5 (table1.csv): N<50 : 26.5% N<100: 57.1% N<200: 80.9%

For all the stars selected as members P>0.01 (members.csv): N<50 : 7.7% (115) N<100: 25.7% (380) N<200: 53.1% (787) N<300: 69.6% (1031)

Mean = 294 stars per cluster (435833/1481) Median = 184 stars per cluster Max = 3646 stars per cluster (NGC_7789) Min = 14 stars per cluster (DBSB_21)

• Cantat-Gaudin et al. (2020)

2017 clusters (234129 stars with P>0.7); parameters for 1867.

Mean = 125 stars per cluster (234129/1867)

Synthetic data

Tested 4 sets of synthetic King profiles to test SVD vs ASteCA with the following parameters:

0. outl_perc=(10, 25); r_max_outl=1.5
1. outl_perc=(5, 20) ; r_max_outl=1.
2. outl_perc=(5, 10) ; r_max_outl=1.
3. outl_perc=5 ; r_max_outl=1.

For all the sets these parameters are fixed: N_memb=200; ell_min=0.2, ell_max=0.8; CI=0.5.

Results of this analysis: not even in the most favorable run (3) is the SVD method able to match the performance of ASteCA for either of the fitted parameters.

Input data

Sol 1613 processed with rad=10*r_50, and the following filters applied:

0. Only for frames with >1000 stars
1. Plx filter: +/- 0.25 (using CG20 mean values)
2. PMs filter: +/- 1 (using CG20 mean values)
3. Center (lon, lat) in ASteCA fixed to CG20 values
4. Auto ASteCA radius