INSIGHTS training event at Nikhef
→
Europe/Zurich
H331 (Nikhef)
H331
Nikhef
Science Park 105,
1098XG Amsterdam
Description
This series of lectures will focus on statistical model building, mathematical models that describe the distributions of observed data in terms of a hypothesis/theory with optional parameters, and the use of standard statistical inference techniques on those models.
The course is structured such that the level of difficulty gradually increases over the duration of the course, as the models are made gradually made more realistic and complex. At each stage ample attention is given to practical and computational aspects of model building, alternated with excursions into statistical theory.
The outline of the course is as follows
Day 1
-
Model building 1 - Simple Poisson Models
-
Statistical methods 1 - Hypothesis testing, p-values, odds ratios (demonstrated on simple Poisson counting experiments)
-
Model building 2 - Modelling distributions - template based models or analytical models
-
Statistical methods 2 - Adapting statistical methods to use with distributions: test statistics as ordering principle, likelihood ratios, contrast with Bayesian methods, the likelihood principle. Practical aspects of toy MC sampling
Day 2
-
Model building 3 - Models with parameters I - analytical parametric models, template morphing approach for histogram-based models
-
Statistical methods 3 - Inference with parameters: maximum likelihood, confidence intervals, upper limits, likelihood ratio and asymptotic formulae
-
Model building 4 - Models with parameters II - simultaneous fits, representing external information as subsidiary measurements (‘profile likelihood fits’)
-
Statistical methods 4 - Parameters of interest vs nuisance parameters, dealing with nuisance parameters in inference methods
Day 3
-
Model building 5 - Diagnostics (understanding MINUIT, fit stability and convergence) and Validation (understanding your fit, overconstraining parameters, 2-point systematics etc)
-
Model building 6 - Advanced model building techniques: analytical lagrangian morphing, matrix element techniques
The course will take place from Wednesday Dec 12 through Friday Dec 14 with 3 lectures in the morning, following by a hands-on session in each afternoon to practice the techniques discussed using ROOT-based exercises.
Practicalities:
This lectures series is jointly offered to Nikhef PhD students and INSIGHTS PhD students. INSIGHTS students should register for this event on this indico page.
Students should bring a laptop for the hands-on session (further detailed information will be provided)
Venue
All lectures will take place in room H331 at Nikhef, Science Park 105, Amsterdam.
Complimentary coffee, tea and refreshments are offered during the lectures. Lunch can be enjoyed either at the Science Park cafetaria, in the same building, or at restaurants nearby (Polder, Maslov)
Travel
Amsterdam Airport (Schiphol) is well (and often cheaply) connected with most European airports
Nikhef is easily reached by train from Schiphol airport. Travel from Schiphol Airport to Amsterdam Central station by train and take from there a train that stops at Amsterdam Science Park. From there, Nikhef is a few minutes walking. Accurate public transport information is readily accessible on e.g. Google Maps. Prices for public transport in and around Amsterdam are typically of the order of 3-5 euros per trip. Taxis are quite expensive (count on e40 from airport to Nikhef) and provide limited added value.
Accomodation
Amsterdam is a popular tourist destination, which is reflected in hotel prices and availability. Decent hotels in the vicinity of Nikhef will cost around e100/night. The Nikhef secretariat can assist you, if desired, with hotel bookings (we do get better rates on some hotels than those publicly advertised). If wish to make use of this service, please contact Ms Joan Berger (jberger@nikhef.nl) and indicate that you are in INSIGHTS student wishing to attend the statistics lectures. In any case, aim to make hotel
bookings (with or without assistence) as soon as possible.
bookings (with or without assistence) as soon as possible.