Ruprecht Karls Universität Heidelberg

GRK 2149 workshop: Statistical Methods in Particle Physics

Lectures for the Retreat of the Research Training Group (Graduiertenkolleg) 2149 "Strong and Weak Interactions" - 9-10 September 2020


  • Basics
  • Maximum likelihood method
  • Least Squares
  • Hypothesis tests and Goodness-of-fit
  • Systematic uncertainties
  • Decision trees

    We'll use python 3 in the course. Basic knowdegle of the language is useful for this course. We'll work a lot with jupyter notebooks. A nice summary of important python commands is available on the website of the Stanford lecture CS231n.

    Here you find the slides of the lecture.



    1. Construct a Bayesian credible interval (html, notebook) ⧖ [solution: html, notebook]
    2. Unbinned maximum likelihood fit (double exponential decay) (html, notebook) ⧖⧖ [solution: html, notebook]
    3. The lighthouse problem: another unbinned maximum likelihood fit (html, notebook) ⧖⧖ [solution: html, notebook]
    4. Unbinned maximum likelihood fit with Gaussian constraint on a parameter (html, notebook) ⧖ [solution: html, notebook]
    5. Linear least squares and error propgation (html, notebook) ⧖⧖ [solution: html, notebook]
    6. Simultaneous least-squares fit to several data sets (blast-wave fit to particle spectra) (html, notebook) ⧖⧖⧖ [solution: html, notebook]
    7. Kolmogorov-Smirnov test (html, notebook) ⧖ [solution: html, notebook]
    8. Significance of a peak above background (html, notebook) ⧖⧖⧖ [solution: html, notebook]
    9. Least-squares fit with external Gaussian constraint (html, notebook) ⧖ [solution: html, notebook]
    10. Separation of gamma and hadron showers measured with the MAGIC Cherenkov telescope using a boosted decision tree and a random forest (html, notebook) ⧖⧖ [solution: html, notebook]
    ⧖ = quick, ⧖⧖ = intermediate, ⧖⧖⧖ = takes a bit longer

    Download zipped tar file with data and jupyter notebooks (unpack with tar xvzf notebooks_and_data.tgz).
  • Webmaster:
    EDV Abteilung