Simulation of Distributions and Monte-Carlo Methods

Seminar im SoSe 2008

Neu + wichtig

• Nothing so far...

Inhalt

Computational Biology extensively uses statistics and probability theory to accept/reject hypotheses, compute significance values as well as confidence intervals. The central concept is the random variable with its main characteristic: the distribution. Some distributions can be computed analytically, others are only accessible by simulations. Furthermore, visualization of distributions can be nicely used to support understanding of many concepts of probability theory and statistics. Hence, we will discuss applications in computational biology involving distributions, compute/simulate the distributions, and (most importantly) use the visualization to understand concepts of statistics and probability theory. Furthermore, we will approach distributions which are infeasible to simulate. In this case, we will use Monte-Carlo methods (i.e. importance sampling) for the computation.

This course is for students who are interested in probability theory and statistics. We assume that everybody has successfully finished the course for statistics (Statistik 1 in the 3rd semester). Attendance of the advanced course of statistics (Multivariate Statistik or Statistik 2) is advantageous but not required. The seminar is obligatory for students attending the corresponding Softwarepraktikum.

Scheinkriterien

• ca. 60-minütiger Seminarvortrag (50%)
• 1-seitiges Handout (10%)
• Seminarbeteiligung (40%)
• Von jedem Teilnehmer wird erwartet, dass er vor jedem Seminartermin den/die entsprechenden Artikel gelesen hat. Pro Termin sind 3 Fragen schriftlich an pape'at'molgen.mpg.de zu senden.

Links

Homepage der Abteilung Computational Molecular Biology am Max-Plack-Institut für Molekulare Genetik.

Zum Studiengang Bioinformatik an der FU Berlin.