Skip to content

Small, but perhaps important fixes to the likelihood function

Gregory Ashton requested to merge fixing-the-likelihood-maybe into master
  • In the likelihood - the signal_IFO now sums only over the polarizations
  • In the likelihood - return only the real part
  • Adds a function to the basic tutorial to plot the loglikelihood over a range of search frequencies
  • Updates model -> frequency_domain_strain (as suggested by @colm.talbot in #3 (closed)

Could someone check the changes to the likelihood are correct? Without these changes, the output doesn't make sense. With these changes the likelihood (as a function of frequency) has the typical sinc^2 form we should expect.

To check this, run the BasicTutorial.py (not the notebook). The second figure produced is the likelihood evaluated over a range of frequencies.

cc @all

Merge request reports