Thursday, July 29, 2010

Continuing Likelihoods

I thought perhaps combining v1 and the new version would produce a better selection algorithm, but this doesn't improve things.

The field l_ratio is computed the following way:

;Calculate lratio
eps = 1e-30
bossqsolike = total(oldlikelihood.l_qso_z[2:18],1) + total(likelihood.L_QSO_Z[start:stop],1) ;quasars between redshift 2.1 and 3.5
qsolcut = where(alog10(bossqsolike) LT -9.0)
den = total(targets.L_EVERYTHING_ARRAY[0:4],1) + total(likelihood.L_QSO_Z[0:44],1) + total(oldlikelihood.l_qso_z,1) + eps
num = bossqsolike + eps
NEWRATIO = num/den
NEWRATIO[qsolcut] = 0 ; eliminate objects with low L_QSO value
likelihood.l_ratio = newratio

Targets/dec^2------v1 threshold----new threshold
-----0------------1.00000--------1.00000
-----10-----------0.749085-------0.784113
-----20-----------0.490932-------0.530647
-----30-----------0.347334-------0.401703
-----40-----------0.253020-------0.311692
-----50-----------0.198630-------0.247766
-----60-----------0.160074-------0.201979
-----70-----------0.128912-------0.171997
-----80-----------0.107913-------0.145289
-----90-----------0.0917588------0.126122


Targets/dec^2--# Targeted QSOs:----v1----new version
-----0------------------------1---------1
-----10----------------------443-------393
-----20----------------------656-------630
-----30----------------------781-------765
-----40----------------------897-------877
-----50----------------------965-------968
-----60---------------------1035-------1030
-----70---------------------1089-------1070
-----80---------------------1123-------1111
-----90---------------------1157-------1151


This doesn't improve things much, and performs worst at 20 targets per square degree (core).

If you look at color-color plots of what was targeted by both version (white), with version 1 (cyan) versus the new version (magenta):


We do seem to be better at targeting higher redshift QSOs with the new method, and missing more lower redshift QSOs, so perhaps this means we should use this method even though it doesn't improve our absolute numbers:

No comments:

Post a Comment