E18/ENE Seminar

Developing Statistical Models: A Demonstration on Partial-Wave Analysis

by Florian Kaspar (TUM E18)

Europe/Berlin
E18 Seminar Room

E18 Seminar Room

Description

When analyzing data, we often face situations in which we have to develop a statistical model to describe the data. While we might be lucky in many cases and the models are simple and the data is abundant, there are also situations in which the model complexity relative to the available data causes the inference to suffer from ambiguous solutions and statistical noise.

An example for the latter case is partial-wave analysis (PWA). At high-energy experiments, we use this tool to gain access to short-lived resonant states. For these analyses, the model choice is a major contributor to the systematic uncertainty. Using PWA as a guiding example, I will demonstrate different approaches on how to tackle the model building step. I will use simulated data to develop and test procedures that are then applied to real data of different experiments.