Séminaire/Seminar Galaxies |
| « Cosmology with a method of iterative smoothing » |
|
Hanwool Koo |
| Iterative Smoothing Method is a non-parametric regression approach for reconstruction of expansion history of the Universe and properties of dark energy. We can apply this method in various aspects of cosmology and astrophysics with no cosmological model assumption. We have worked on constraining SN Ia light-curve hyperparameters, searching for features in the data which can be a hint for systematics or new physics, parameter estimation and model selection. Our most recent application of the method is comparing our novel frequentist approach for model selection with conventional Bayesian evidence model selection. We show that when none of the candidate models (proposed to fit a data) is the true model, our novel frequentist approach can rule out all candidates while conventional Bayesian approach selects the least incorrect model. For more practical application of the method, our ongoing task is developing method of iterative smoothing for reconstructing the expansion history using combination of data with different natures. The method using combination of Type Ia supernova and BAO data is in preparation. We expect that the new method can be applied for cosmological studies using the real data from present and forthcoming surveys including DESI, Rubin, Roman, etc. |
|
jeudi 13 octobre 2022 - 11:30 Salle 281 Institut d'Astrophysique de Paris |
| Page web du séminaire / Seminar's webpage |