Multiifo coincident rate estimation
This page details the testing done to the coinc_rate.multiifo_noise_coinc_rate function to check that noise rates are being calculated correctly
Note that there is some 'blocking' (grouping of templates with similar duration) in order to obtain better statistics for the triple coincident events
This 'overestimate' is the predicted rate divided by the true rate, there may be a small bias due to the true rate being an integer
This rate can be used as a benchmark/constant standard for scaling rates from other ifo combinations and/or data periods to obtain numbers of order 1. Here we look at a few recent runs representative of 2-OGC or O3 production settings.
- directory
/home/tdent/o2/analysis/o3dev/a22/multiifo_20190406_klsgv/output
mean actual noise coinc rate per template = 4.5e-7 Hz
- O2 a22: 15 splitbanks, random-sort, keep-loudest-interval=1, keep-loudest-num=100, coinc threshold 0.005
directory: /home/tdent/o2/analysis/o3dev/a22/multiifo_20190419_expfit_decim_shuffle2/output
mean expected rate per template in different coincidences:
HLV coincidences is 1.101e-10 Hz
HL coincidences is 4.563e-07 Hz
LV coincidences is 9.555e-07 Hz
HV coincidences is 9.750e-07 Hz
- O2 a20: 15 splitbanks, random-sort keep-loudest-interval=1, keep-loudest-num=100, coinc threshold 0.005
directory: /home/tdent/o2/analysis/o3dev/a20/multiifo_20190419_expfit_decim_shuffle/output
mean expected rate per template in different coincidences:
HLV coincidences is 1.095e-10 Hz
HL coincidences is 4.550e-07 Hz
LV coincidences is 9.524e-07 Hz
HV coincidences is 9.712e-07 Hz
- O3 a1: 20 splitbanks, mchirp-sort, keep-loudest-interval=1.072, keep-loudest-num=100, minimal DQ, coinc threshold 0.002
directory: /home/tdent/o3/analysis/c00_clean/hlv/a1/initial/output
Mean HL coincidences per template: 4.985e-07