Why is there a need for Bayesian parameter estimate analysis if Match filtering can detect GW signals? Is Match filtering used only for detection then Bayesian Inference techniques are used for a better complete description of the event?

I apologize in advance if my question is a bit silly and basic! I am currently undergoing workshop #5 and must say it has been highly informative to a beginner like me, I am currently an undergrad studying Physics.

From what I recall watching the talks given by the experts, Match filtering uses a large template bank of simulated gravitational waveforms for different parameters which are then used to detect the presence of similar waveforms in unknown signals (given by the Signal to Noise ratio across the signal)

If a particular unknown signal matches with one of the templates (with a high enough SNR), would that not infer the occurrence of an event similar to that of the template? So why is there a need for Bayesian parameter estimation techniques, (also from what I recall there can be 15 parameters that can describe an event, are these all of the 15 considered to generate the waveforms used in the template bank)

Also, my General Relativity knowledge is not too well but say at a given time we consider the superimposed effects of all the gravitational waves Impacting Earth, should that affect the distances and time we experience?

Thanking in advance

2 Likes

@herwins Thank you for your question!

More or less: Yes! Matched filtering is used for detection, and Bayesian parameter estimation is used for a more complete description of the source. Here are a few differences:

  • To make a search with a template bank possible computationally, we typically search only over 3 parameters (m1, m2, and effective spin). So, we don’t get measurements for all 15 parameters from the matched filter search.

  • Because the parameter estimation code uses a Bayesian statistics approach, it measures not only approximate values for the parameters, but also the uncertainty (or credible intervals) for each parameter, as well as correlations between different parameters. That is, PE code measures a 15 dimensional probability distribution to describe the source, not just a single set of values.

Good luck!

2 Likes

@herwins For the 2nd question, the short answer is: Yes! Gravitational waves passing earth create small distortions in spacetime.

Of course, the impacts of passing gravitational waves are very small, which is why we don’t notice them every day (e.g. I’ve never missed a train because a gravitational wave made me late). However, the superimposed effects of passing gravitational waves should be present all the time. There are several possible sources of these, but the one we are likely to measure first is the superposition of many gravitational waves from many distant binary black holes.

With LIGO, Virgo, and KAGRA, the Stochastic Working Group tries to measure these signals, by looking for a correlated signals across multiple observatory sites.

Also, Pulsar Timing Array projects, such as NANOGrav, are working to measure a stochastic signal from very distant super-massive binary black holes in the centers of galaxies. You can read about their work on the NANOGrav website.

Good luck!

1 Like

Physics and the world around is incredibly beautiful, work of you scientists is applauding, Thank you for your efforts and dedication. I shall go over your shared resources, there is still more for me to get my head around.

To all that is left to find out and explore and solve I wish us all good luck!

1 Like