15.04.2014 Views

Statistics, Data Analysis, and Simulation SS 2013

Statistics, Data Analysis, and Simulation SS 2013

Statistics, Data Analysis, and Simulation SS 2013

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong><br />

<strong>SS</strong> <strong>2013</strong><br />

08.128.730 Statistik, Datenanalyse und <strong>Simulation</strong><br />

Dr. Michael O. Distler<br />

<br />

Mainz, 16. April <strong>2013</strong><br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


The Mainz Microtron<br />

1.6 GeV cw electron beam<br />

100 µA unpolarised beam<br />

30 µA beam @ 80 % polarization<br />

Energy stability δE/E = 10 −6<br />

> 5500 h/yr available for experiments<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


The 3 spectrometer facility<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Learning goals <strong>and</strong> objectives<br />

Students will gain a basic knowledge of<br />

statistics, simulation techniques, <strong>and</strong> numerical methods<br />

(algorithms)<br />

Problem:<br />

determine meaningful <strong>and</strong> significant information from<br />

experimental (empirical) data<br />

efficient (economic) data analysis<br />

Application of this knowledge in the data analysis<br />

Probability of events<br />

Uncertainty of a measured quantity<br />

Significance of a measurement (discovery)<br />

Decision rules for (testing of) model hypotheses<br />

Determine (estimate) the best values of parameters<br />

<strong>Simulation</strong> of complex processes<br />

Unfolding, factor analysis, pattern recognition, . . .<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Contents<br />

Introductory Remarks<br />

Statistic:<br />

probability, distributions, discrete distributions, special<br />

continuous distributions, theorems, sampling, multidimensional<br />

distributions<br />

Monte Carlo methods:<br />

r<strong>and</strong>om number generators, Monte Carlo integration<br />

Parameter estimation:<br />

the method of maximum likelihood, variance of the estimators<br />

The method of least squares:<br />

linear least squares, properties, generalized least squares<br />

Hypothesis testing<br />

Advanced topics:<br />

unfolding, factor analysis, pattern recognition, . . .<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Literature<br />

V. Blobel, E. Lohrmann: Statistische und numerische<br />

Methoden der Datenanalyse, Teubner Verlag (1998),<br />

available as http://www.desy.de/∼blobel/eBuch.pdf<br />

S. Br<strong>and</strong>t: Datenanalyse, BI Wissenschaftsverlag (1999)<br />

Philip R. Bevington: <strong>Data</strong> Reduction <strong>and</strong> Error <strong>Analysis</strong> for<br />

the Physical Sciences, McGraw-Hill (1969)<br />

R.J. Barlow: <strong>Statistics</strong>, John Wiley & Sons (1993)<br />

G. Cowan: Statistical <strong>Data</strong> <strong>Analysis</strong>, Oxford University<br />

Press (1998)<br />

W.T. Eadie et al.: Statistical Methods in Experimental<br />

Physics, North Holl<strong>and</strong> Publishing Company<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Administration<br />

Home page: http://wwwa1.kph.uni-mainz.de/<br />

Vorlesungen/<strong>SS</strong>13/Statistik/<br />

Lecture time: Tuesday, 12:15-13:00,<br />

Thursday, 8:30 - 10:00.<br />

<strong>and</strong> place: seminar room 1, Institut für Kernphysik<br />

Problem solving time: Tuesday, 13:15 - 14:00<br />

<strong>and</strong> place:<br />

seminar room 1, Institut für Kernphysik<br />

“Studienleistung”:<br />

regular <strong>and</strong> active participation in the practical exercises,<br />

solve at least 10 problem sets (out of 12),<br />

achieve > 50% total score.<br />

“Prüfungsleistung”:<br />

oral examination (30 min) e.g. “Vertiefungsvorlesung<br />

(Master)”, “benoteter Schein”<br />

none e.g. “Spezialvorlesung (Master)”, “unbenoteter Schein<br />

(Diplom)”<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Problem solving classes<br />

“classical” problem sets (comprehension <strong>and</strong> calculation)<br />

“computer based” problems (put the theory into practice)<br />

problems will be h<strong>and</strong>ed out after the Thursday lecture<br />

(<strong>and</strong> will be available online)<br />

due date is the following Thursday, 10am.<br />

Dr. Sören Schlimme will grade the answers <strong>and</strong> discuss<br />

the solutions during the problem solving classes.<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Examples<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Examples<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Examples<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Introductory Remarks<br />

<strong>Data</strong> analysis is used in many scientific fields,<br />

but the lecture here is aimed at mainly physicists.<br />

Therefore, a preliminary remark from this view:<br />

Physics is the science of quantifiable observations.<br />

The comparison:<br />

observations ←→ classification scheme<br />

takes place quantitatively, ie it comes to numbers.<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Introductory Remarks<br />

<strong>Data</strong> analysis in nuclear <strong>and</strong> particle physics<br />

Observe events of a certain type<br />

Measure characteristics of each event<br />

Theories predict distributions of these properties up to free<br />

parameters<br />

Some tasks of data analysis:<br />

Estimate (measure) the parameters;<br />

Quantify the uncertainty of the parameter estimates;<br />

Test the extent to which the predictions of a theory are in<br />

agreement with the data.<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Introductory Remarks<br />

Theory: numbers are calculated using a model.<br />

Experiment: numbers are derived from observation.<br />

This raises the issue of consistency<br />

between theory <strong>and</strong> experiment.<br />

What does consistency or agreement mean?<br />

Is there a measure of (non-) agreement?<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Introductory Remarks<br />

Philosophy of Science<br />

Karl R. Popper (* 28. Juli 1902 in Vienna, Austria;<br />

† 17. September 1994 in London, Engl<strong>and</strong>) coined the term<br />

critical rationalism. At the heart of his philosophy of science lies<br />

the account of the logical asymmetry between verification <strong>and</strong><br />

falsifiability. Logik der Forschung, 1934.<br />

−→ Existence of a true value<br />

of measured quantities <strong>and</strong> derived values.<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>


Recipe: How to do science?<br />

observation<br />

(of nature)<br />

hypothesis<br />

(model)<br />

deduction<br />

experiment<br />

(reproducible)<br />

prediction<br />

(falsifiable)<br />

? Truth ?<br />

Important philosophical questions:<br />

How to gain knowledge?<br />

What is truth?<br />

What is a scientific theory?<br />

Dr. Michael O. Distler <strong>Statistics</strong>, <strong>Data</strong> <strong>Analysis</strong>, <strong>and</strong> <strong>Simulation</strong> <strong>SS</strong> <strong>2013</strong>

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!