Astrophysics (Index) | About |

**Gini/M20** (or **Gini-M20**) is a method of classifying galaxies
both regarding morphology and whether they are recent mergers, using
pixel data. It can be done algorithmically, thus is suitable for
application to survey data, and classifying whole
populations of galaxies.

It is based upon a graph of two statistics: **Gini** (or **G**) and
**M20** (or **M _{20}**) applied to the flux measure at each pixel:

- Sufficiently high Gini and M20 implies a galaxy merger.
- Other combinations, depending upon how high Gini and M20 are imply various galaxy classifications.

*Gini* (G) is a statistic regarding the distribution of flux among
the pixels, varying from G = 0 meaning each pixel received the same
flux, to G = 1 meaning all the flux was to a single pixel. (The
statistic is borrowed from economics where it is used to describe
wealth distribution of populations, e.g., few rich with lots
of wealth, or a wider spread.)

*M20* is a statistic regarding how the brightness is positioned across
the galaxy: in particular how far the bright pixels are from the
center. The statistic is the log of a ratio: a "moment" regarding
a bright group of pixels versus the same moment for all the galaxy's
pixels. The "bright group" is the pixels receiving the most flux,
the set cut off when their sum constitutes 20% of all the flux.
The "moment" is the sum of products for each pixel, the product
being pixel's flux received multiplied by the square of its distance
from the center of the galaxy, presumably a *center-of-mass* type
center based on flux received, pixel by pixel.

http://en.wikipedia.org/wiki/Gini_coefficient

https://ui.adsabs.harvard.edu/abs/2004AJ....128..163L/abstract

http://astrostatistics.psu.edu/su11scma5/lectures/freeman_scmav_poster.pdf

https://lavinia.as.arizona.edu/~gbesla/ASTR_520_files/Lecture16_GalaxyStructure.pdf