Commissioning Quasars

I am trying to answer these questions about the likelihood targets in the commissioning data. Today I took the collated targets for the QSO commissioning region from the wiki (here) and used spherematch to align it with the truth table that Nic Ross has provided me (based on visually inspected spectra). For truth table see the Nov 20th email from Ross with subject line BOSS Quasars "Truth Table." Both are also on riemann in the directory:

/home/jessica/boss/BOSS_QSOs_First14plates.fits
/home/jessica/boss/bosstarget-qso-comm-collate.fits

Target bitmasks from the wiki:

# restrictive qso selection
maskbits BOSS_TARGET1 10 QSO_CORE
# permissive qso selection
maskbits BOSS_TARGET1 11 QSO_BONUS
# known qso between [2.2,9.99]
maskbits BOSS_TARGET1 12 QSO_KNOWN_MIDZ
# known qso outside of miz range. never target
maskbits BOSS_TARGET1 13 QSO_KNOWN_LOHIZ
# Neural Net that match to sweeps/pass cuts
maskbits BOSS_TARGET1 14 QSO_NN
# UKIDSS stars that match sweeps/pass flag cuts
maskbits BOSS_TARGET1 15 QSO_UKIDSS
# kde targets from the stripe82 coadd
maskbits BOSS_TARGET1 16 QSO_KDE_COADD
# likelihood method
maskbits BOSS_TARGET1 17 QSO_LIKE

These are bitmasks, so in IDL to select items that are only selected by the likelihood method we would do the following:
LIKEONLY = where(truthmatch.BOSS_TARGET1 EQ ishft(long(1),17))

To select items which are selected by the likelihood (and others):
LIKE = where((truthmatch.BOSS_TARGET1 AND ishft(long(1),17)) GT 0)

To select items which are not selected by the likelihood (selected by others):
NOTLIKE = where((truthmatch.BOSS_TARGET1 AND ishft(long(1),17)) EQ 0)

To select objects with high redshift:
HIGHZ = where(truthmatch.Z_PERSON GT 3)

Below is a color-color plot of the above selections on top of the templates used for the likelihood selection. Red is everything. Green is QSOs. Magenta are objects that the likelihood method selected (LIKE). Cyan are objects that only the likelihood method selected (LIKEONLY). White are objects that have a high redshift (HIGHZ):

Color-Color (u-g vs g-r)
of likelihood selected objects

Below are objects that are QSOs but were missed by the likelihood (NOTLIKE). Need to figure out why.

Color-Color (u-g vs g-r)
of not-likelihood QSO objects

Back to the Likelihoods

It's been a while since I have worked on the likelihood QSO selection method. With the next deadline for target selection is coming up, it's time to go "Back to the Likelihoods." Note to self: It more time efficient to keep working on something continuously than to not work on it for weeks and then waste a day trying to remember what I was doing.

Where we left off...
Below is a color-color (ug - gr) plot of the final likelihood selection objects for the commissioning data.

The white data points are a random sampling of 20,000 possible objects to target. The red data points are objects who's likelihood ratio is greater than 0.1, where likelihood ratio is defined as:

where L_QSO and L_everything, as described in "A Likely Result" are defined as:

The green data points are objects are objects who's likelihood ratio is greater than 0.1 and L_everything is greater than 10^-6. This is eliminate classification of "fringe" objects that are not close to any objects (and therefore have a small everything likelihood.

The likelihood was then run on the co-added Stripe 82 data and all of the above green objects were submitted as targets for the commissioning data.

What we need to work out...

• Why are the likelihoods so small/large? The likelihoods should be a probability, but we have likelihood's spanning from 0-11.
  
• Why is our completeness and efficiency on the MMT data so poor?
• How does the likeliness compare to QSOs based on variability?
• How well does this method work on single epoch Stripe 82 data versus the co-added images?