You are here: Wiki>DiBiMath Web>StatistikWS13 (09 Apr 2013, HeikeSiebert)Edit

Wilkommen beim Wiki "Statistik", WS 2012/2013

Kontakt
Heike Siebert
Arnimallee 7, Raum 103
Office hours (during lecture time): Wednesday, 11:30-12:30 ( not on February 6 2013 )

Bernhard Renard
Robert Koch-Institut, Nordufer 20, Raum 312
Office hours: after the lecture and by email appointment

Natasa Djurdjevac
Arnimallee 6, Raum 127
Email: djurdjev[at]mi.fu-berlin.de

News

The results for the Nachklausur are online.
You can have a look at your exam on April 24, 18:15 - 18:30 in Arnimallee 7, Room 103.

The results for the exam are online. (If your result is missing, you did not grant us permission to publish your results on the web - please send an email with your exam number)
On February 26, 16:30-17:30, you can come by and have a look at your exam in Arnimallee 7, R103 (Heike Siebert's office). You can also take this opportunity to have a look at the exam if you did not write the exam on Feb. 19. Please send an email if you want to take that opportunity.

Problem set 14 will be a bonus sheet. You can use it to collect some more points if you did not make the 50% on sheets 1-13. The topic will be MCMC so you might want to have a look at the type of problems regardless!

There will be no lecture on Feb 12. However, starting at 9:00 we will give a short review of the lecture content and answer questions concerning the exam. If you have any specific questions to some lecture topic please email us your question in advance, so we can deal with it efficiently.

Registration at Campus Management should be possible now. Please register!

Exam

Nachklausur: April 9, 18:00-19:30, A6 SR 031
As announced, we are not allowed to offer a full 'Freischuss' (the possibility to retake the exam and choose the better grade from either exam). If you passed the exam on Feb 19 but are not happy with your grade, you can contact us and ask for your exam to count as a failing grade and retake the exam on April 9. Be sure to understand, however, that such a decision to have the exam evaluated as a failing grade is not retractable! The result of the April exam will then count, regardless of whether it is better or worse than your original result. So please be sure of your decision.

Exam day is February 19th, 10 am, Arnimallee 3, Room 001

No aids are allowed. You just need a pen, we'll provide paper.

In Bioinformatics there is no "Freischuss". For those who did not pass, there will be another exam (probably on April 9 18-20).

General Informationen

Lecturers: Bernhard Renard, Heike Siebert
SWS: 2

Exercises: Natasa Djurdjevac
SWS: 2

ECTS: 6
Sprache: English/Deutsch

Dates and Locations

Lecture:
8:30 - 10:00 on Tuesdays in room 031, Arnimallee 6
Start: 16.10.

Exercises:
12:15 - 13:45 on Wednesdays in room 119, Arnimallee 3
14:15 - 15:45 on Wednesdays in room 130, Arnimallee 3 (backbuilding)
Start: 24.10.

Topics

B. Renard:
Computational Statistics and Statistical Learning
H. Siebert:
Mathematical background for Markov chains and related topics.

Requirements

Exercises are mandatory, problem sets will be posted on this website on a weekly basis and are to be handed in at the Tuesday lecture. At least 50% of all graded problems need to be passed for a successfully participation. Attending the lecture is highly recommended. A 90 minute final examination determines the final grade.

Literature

Hastie, Tibshirani & Friedman. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, 2009. http://www-stat.stanford.edu/~tibs/ElemStatLearn/download.html

Volker Schmidt. Markov Chains and Monte-Carlo Simulation, Lecture Notes University Ulm, 2010. Available here.
Pierre Bremaud. Markov Chains, Gibbs Fields, Mote Carlo Simulation, and Queues. Springer 1999. Available here.
Ehrhard Behrends, Introduction to Markov Chains (with Special Emphasis on Rapid Mixing), Vieweg, 1999.

Lecture Notes

Part 1
01 Introduction: Slides
02 Multiple Regression: Slides Lecture Notes
03 Kernel Density Estimation: Slides Lecture Notes
04 Non-Parametric Regression: Slides Lecture Notes
05 Model Evaluation: Slides Lecture Notes
06 Regularization: Slides Lecture Notes
07 Classification Trees: Slides Lecture Notes
08 Ensemble Methods: Slides Lecture Notes
09 Normalization: Slides
10 Summary: Slides

Part 2
01 Basic Definitions (References: Schmidt Section 2.1.1 and Ex 1 of 2.1.2, Bremaud Chapter 2.1 ):
stochastic process, homogeneous Markov chain, transition matrix, transition graph
02 Canonical representation, n-step transition and stationarity; notes
03 Communication and absorption; notes
04 (A)Periodicity; notes
05 Ergodicity and stationarity; notes
06 Reversibility; Introduction to Markov Chain Monte Carlo; notes
07 Markov Chain Monte Carlo: Metropolis-Hastings-Algorithm; notes

Exercises

Exercise 0 R script (not to be handed in)
Exercise 1 (due Oct 30 in the lecture) solution
Exercise 2 protein data tumor data (due Nov 06 in the lecture) solution
Exercise 3 ehec data (due Nov 13 in the lecture) solution
Exercise 4 chlamydia psittaci genome chlamydia trachomatis genome (due Nov 20 in the lecture) solution
Exercise 5 spectrum models (due Nov 27 in the lecture) solution
Exercise 6 data (due Dec 04 in the lecture) solution
Exercise 7 (due Dec 11 in the lecture)
Exercise 8 (due Dec 18 in the lecture)
Exercise 9 (due Jan 8 in the lecture)
Exercise 10 (due Jan 15 in the lecture)
Exercise 11 (due Jan 22 in the lecture)
Exercise 12 (due Jan 29 in the lecture)
Exercise 13 (due Feb 5 in the lecture)
Exercise 14 (bonus sheet, due Feb 12 in the lecture)
Topic revision: r46 - 09 Apr 2013, HeikeSiebert
 
  • Printable version of this topic (p) Printable version of this topic (p)