Almost all statistical analyses begin with the words: If the data comes from an XYZ distribution, then ..., so an obvious question is: how can we be sure that a certain data set has been generated by a given probability distribution?
In this talk I will discuss some of the many tests that have been developed for this question, starting with the grandfather of all goodness-of fit tests, Pearson's chi square. Other tests include Kolmogorov-Smirnov, Anderson-Darling, Zhang's likelihood tests, Neyman's smooth tests etc. I will also talk about gof testing in higher dimensions and issue of the curse of dimensionality. I will discuss a number of power studies that show a simple truth about goodness-of-fit testing: one size does not fit all!
The seminar will be done remote only.
M. Girone, M. Elsing, L. Moneta, M. Pierini
Event co-organised with the PHYSTAT Committee