Date
December 05-09, 2011.
Topics
Introduction into Statistical Methods
Lecturers
Helge Voss, MPI für Kernphysik, Heidelberg
Practical Information
Begin/End
The power week starts with an introduction session on Monday morning and will close on Friday late afternoon.
Late arrivals
If you will arrive late please contact the coordinator in advance.
Venue
The power week will be held at Theologisches Seminar Herborn, Nassaustraße 36, 35745 Herborn.
Travel
Please arrange for individual travel to the location. We enourage to use car pools.
Accomodation
Accomodation is in double rooms.
Food
Special food requirements (vegetarian, other) will be arranged on the first day.
Full board is included in the lecture week.
What to bring
The power week will be very informal, so only casual attire necessary.
Linen and towels will be provided. Please bring sufficient amounts of paper
and a laptop if possible.
Internet
Wireless LAN is available.
Expenses
All basic expenses are covered by HGS-HIRe.
You only have to pay for local expenses (drinks,etc.). All meals are included.
Insurance
Though this trip will not generate any costs for your supervisor or group please
do not forget to file a trip request so you are covered by insurance during the
lecture week.
Program
The program of this statistics power week will be as follows:
The lecture series will cover a general introduction into statistics,
explaining the basic concepts of probability theory and giving an
explanation of the differences between classical(frequentist) and
Bayesian statistics. We will introduce random variables, probability
densities, expectation values and how this relates to our daily
measurements. This will be continued by discussing statistical tests,
hypothesis testing and parameter estimation. This will lead us to the
discussion of confidence levels, limit and how to incorporate
systematic uncertainties.
The second half of the week will be devoted to multivariate
classification and regression techniques as an application of
statistical tools. After introducing the general idea of multivariate
classification algorithms, we will get familiar with
(Multi-dimensional) Likelihoods, Linear Discriminants, Neural Networks
as well as more recent developments of machine-learning tools such as
Support Vector Machines and Boosted Decision Trees.
While the whole lecture series with accompanying exercises is meant to
be self contained without explicit prior knowledge about these
statistical concepts, a decent knowledge of C++ and basics concepts
of the ROOT data analysis framework (root.cern.ch) will be needed
in order to follow the exercises.
What to prepare
A basic knowledge of the ROOT programming language and or C++ is recommended for the lecture week. Below are two tutorials that give a quick introduction.
Tutorial I
Tutorial II
It is necessary that you bring a laptop with a working ROOT installation (please check before the Power-Week how to set the compiler up, it will save a lot of time during the week.)
If you do not have a laptop available or have trouble setting up ROOT, please contact us.
Participants
- Jordi Albert-Fort
- Martha Liliana Cortes
- Dennis Döring
- Marcel Heine
- Jan Kopfer
- Matthias Kretz
- Igor Kulakov
- Vladimir Lavrik
- Daniel Löb
- Mareike Müller
- Valerii Panin
- Christopher Pinke
- Christian Schäfer
- Iurii Sorokin
- Tian Zhang
Accomodation