[Skip navigation]

[På svenska]

Overview

Lectures:

Mondays 8-10, MH:309A
Thursdays 13-15, MH:309A
Except October 1:st (MA:3) and October 12:th (MH:362D)
The first lecture is given on Monday 31 August. See also the schedule below.

Computer exercises:

Tuesdays 15-17, MH:230
or
Wednesdays 15-17, MH:230
With the first lab on September 8.
For those of you who are new to matlab:
Introduktion till Matlab (in Swedish)
Introduction to Matlab (Mathworks)
Matlab to R reference

Office hours:

Mondays 13-15, MH:245
Thursdays 15-17, MH:245

Examination

Six computer exercises, each requiring a very short report (no longer than one page). Three home assignments/projects, with oral examination after the last project (focus will be in the last project, but all three projects may be discussed).
The assignments will be handed out during the 2:nd, 4:th and 6:th course week.

v2: 1) Simulation and Monte-Carlo integration, (pdf)
Handed out 2009-09-07, due 2009-09-24.

v4: 2) MCMC and Bayesian inference, (pdf)
Data: coal_mine.txt and challenger.txt)
Handed out 2009-09-21, due 2009-10-08.
Several of the projects have now been graded (I'm still working on a few that handed in late). If you want feedback the projects are available in my office.

v6: 3) Bootstrap, (pdf)
Estimation function: est_negbin.m
Data: austria.txt
denmark.txt
france.txt
germany.txt
greece.txt
holland.txt
italy.txt
spain.txt
sweden.txt
switzerland.txt
Handed out 2009-10-06, due 24 hours before your oral exam.

Tips:

• Example 8.13 provides one way (there are others) of testing different models against each other.
• The density for each observation is negative binomial with r=1/alpha and p = 1/(1+alpha*exp(...))
• Permutation test are probably not a good idea.
• Matlab's nbinrnd can be used to simulate from a negative binomial.
• The dataset is large, consequently the code may well take a couple of hours to run.
• To find a fast computerlab check the list of hardware in the E-building and Maths-building. In general, the newer installations are faster.

Oral exam

Times for the oral exam are

• 3/11 15:15-15:45 (0 places left) MH:329
• 4/11 15:15-15:45 (0 places left) MH:329
• 5/11 15:15-15:45 (0 places left) MH:329

If you have not signed up yet please contact the lecturer (additional times may be considered given sufficient interest).

Schedule

Updated 2009-09-17.

Day				Lectures (chapters in the book)	Computer Ex.
v1/36	Mon	31/8	L1	Introduction		pdf
	Thu	3/9	L2	Random number generation (2-3) Ziggurat algorithm		pdf
v2/37	Mon	7/9	L3	Monte Carlo-integration (4)		pdf
	Tue	8/9	C1		pdf
	Wed	9/9	C1
	Thu	10/9	L4	MCMC (5.1,5.3,5.5-5.6) Equation of state calculations by fast computing machines (1953) Monte Carlo sampling methods using Markov chains and their applications (1970) Extra material about MCMC pdf		pdf
v3/38	Mon	14/9	L5	MCMC (5.2,5.4,5.7) MCMC example, sampling from Gumbel ( example.m)		pdf
	Tue	15/9	C2		pdf
	Wed	16/9	C2
	Thu	17/9	L6	Stochastic modelling and Bayesian inference (6.1,10.1) MCMC example, change point ( MCMC_Exp.m)		pdf
v4/39	Mon	21/9	L7	Bayesian examples, simulation (10.2-10.3,11)		pdf
	Tue	22/9	C3		pdf
	Wed	23/9	C3
	Thu	24/9	L8	Statistical models (6)		pdf
v5/40	Mon	28/9	L9	Bootstrap (7.1-7.3) A leisurely look at the Bootstrap, the Jackknife, and Cross-Validation (1983)		pdf
	Tue	29/9	C4		pdf ph1.txt and ph2.txt
	Wed	30/9	C4
	Thu	1/10	L10	Parametric Bootstrap (7.4-7.5)		pdf
v6/41	Mon	5/10	L11	Permutaion test (8)		pdf
	Tue	6/10	C5		pdf atlantic.txt est_gumbel.m
	Wed	7/10	C5
	Thu	8/10	L12	The EM-algorithm (9) Maximum Likelihood from Incomplete Data via the EM Algorithm (1977)		pdf
v7/42	Mon	12/10	L13	Summary, comments No, the pdf file is not broken. It just contains huge imges.		pdf
	Tue	13/10	C6		Help with project 3
	Wed	14/10	C6		Help with project 3

Literature

M. Sköld, Computer Intensive Statistical Methods and some additional handouts.
The book is available from the Department for Mathematical Statistics.

The above book is the only one needed for the course.
However if you wish to explore other literature some good options are:

Monte Carlo

• Markov Chain Monte Carlo in Practice, Gilks, Richardson and Spiegelhalter, 1996
• Monte Carlo Statistical Methods, Robert and Casella, 2005

Bootstrap

• Bootstrap Methods and Their Application, Davison and Hinkley, 1997
• An Introduction to the Bootstrap, Efron and Tibshirani, 1994

People

Course administrator/lecturer:

Johan Lindström
room: MH:245
phone: 046-222 40 60
e-mail: johanl@maths.lth.se

Computer exercises:

Jonas Wallin
room: MH:237b
e-mail: wallin@maths.lth.se