Program **Links under construction!!!**

**klausUr2SS12**

Ultimas
Noticias

**Check this out:** IDE for R. download free4academia (>500MB):

http://www.revolutionanalytics.com/products/enterprise-productivity.php

**maps 4everything in R here:**

**http://blog.revolution-computing.com/2009/10/geographic-maps-in-****r.html**

**CIPIII free each Tu 16-20! Can we move up from 8-10?**

i have never forgotten this article and come back to
it now and then

http://www.nytimes.com/2008/07/31/opinion/31kristof.html?em

**New EX2 for Monday, 14 ^{th} April**

The nyt on Bayes
Theorem:

Specifically Bayes’s
theorem states (trumpets sound here) that the posterior probability of a
hypothesis is equal to the product of (a) the prior probability of the hypothesis
and (b) the conditional probability of the evidence given the hypothesis,
divided by (c) the probability of the new evidence.

Consider a concrete
example. Assume that you’re presented with three coins, two of them fair and
the other a counterfeit that always lands heads. If you randomly pick one of
the three coins, the probability that it’s the counterfeit is 1 in 3. This is
the prior probability of the hypothesis that the coin is counterfeit. Now after
picking the coin, you flip it three times and observe that it lands heads each
time. Seeing this new evidence that your chosen coin has landed heads three
times in a row, you want to know the revised posterior probability that it is
the counterfeit. The answer to this question, found using Bayes’s theorem
(calculation mercifully omitted), is 4 in 5. You thus revise your probability
estimate of the coin’s being counterfeit upward from 1 in 3 to 4 in 5.

A serious problem arises,
however, when you apply Bayes’s theorem to real life: ……

where: CIP-III (law library 2^{nd} staircase
down to 2^{nd} basement)

when: mo 10.15-11.45, tu 8.30-10.00ßchange!

start: 02/04/12

what: http://www.r-project.org/

Course
description

This course is about
simulation methods in statistics. It is divided into 2 parts:

MCMC (Markov chain
monte carlo methods) and bootstrapping

About the first part:

Simulation methods
use random numbers, so we will start out with random number generation.

MCMC is rooted in
Bayesian statistics (remember Bayes rule?) though it is applied in sampling
based (frequentist) statistics as well.

We will give an
introduction to Bayesian statistics in order to understand MCMC. Though many
results from classical statistics can be obtained as special cases of Bayesian
statistics it will become apparent that this radically different theory can also
be applied to inference problems where sampling based methods fail, for
example: the sample size is too small, there are too many parameters, or
computations would be too complicated.

This course will be
computer intensive. Computations are an indispensible part. We use computations
as heuristic tool to understand or illustrate the theory as well as to illustrate
practical applications of the methods we learn. We use the languages R and WinBugs . No previous knowledge
of programming in these languages is required. There will be programming
exercises weekly. Soon you will be fluently programming in R soon.

Evaluation: The final
grade is computed from weekly programming exercises and a written final exam.

Prerequisites are a
course in basic probability calculus (conditional probabilities, Markov chains)
and in basic statistics.

The course will
consist of 25% bootstrapping and 75% mcmc. If time permits we will have also
have a chapter on the EM-algorithm.

There will be more
details on this outline in the coming days. For questions or suggestions write me a message.