SF2955 Computer intensive methods in mathematical statistics
  • (23/2) The first lecture will be held on 22 March 2017, 8-10, in L51. The course schedule is found here. The material presented during the lectures will be posted on this page after each lecture. Some additional notes that cover partly the contents of the course can be downloaded here. A preliminary course plan is found below, where the assignment program will be updated within soon.
  • (28/3) The course is up and running, all exercises are up and the slides for the first three lectures are up. Please look at the exercises before the exercise class so we can have a more interactive environment.
  • (6/4) The file for Home Assignment 1 has now been uploaded. Your group number can be found here. If you can't find yourself in that list and want to complete the course contact Johan Westerborn at johawes@kth.se. If you find any errors in the assignment or have any questions please don't hesitate to contact Johan Westerborn.
  • (10/4) Home Assignement 1 has been updated to correct some small errors.
  • (20/4) Small update to HA1. Also update the slides from Lecture 7 and 8, there was a small error in the weight updating of the SISR algorithm that is now corrected.
  • (28/4) A LaTeX template for project reviews can be found here. Each project report is assessed anonymously with respect to contents, presentation, evidence, and overall effectiveness. Do not forget to motivate and support your assessment with specific comments concerning the different criteria, at least three comments per criteria; such comments play a critical role in the case of a revision of the project report under consideration. Try to be as clear and specific as possible. Each review should be emailed to the lecturer (johawes@kth.se) by the corresponding group member by Tuesday 9/5, 23:59. Note that the reviews are performed groupwise, i.e., you only need to submit two reports per group, not per student. Of course, if both members of a group feel like submitting separate reports, this is all fine (since it increases the amount of feedback).
  • I've now sent out the reports for review to every group through the e-mail that you submit everything to me. So if you didn't recieve one talk to your groupmember they should have it. If no one of you have it then send me an e-mail.
  • (11/5) The reports has been graded (P/Fx) and you should have recieved your grade in an e-mail. If you recieved the grade Fx please send in a corrected version of the report to johawes@kth.se by May 24, 2017.
  • (8/6) The exam is now marked, and the results have been handed over to the administration. Students who obtained the grade "Fx" should, as soon as possible, get in contact with Johan (johawes@kth.se) in order to obtain information concerning how to complement the exam for passing the same. Full solutions to the exam will be posted in the near future.
  • The re-exam has been graded and you should have recieved your results. Students who obtained the grade "Fx" should get in contact with Johan (johawes@kth.se) as soon as possible.
  • Schedule

    Day Lectures (book chapters) Assignments
    v122/3, 8-10L51L1 Introduction, the Monte Carlo (MC) method (4)
    23/3, 10-12L51L2 MC (cont.), Random number generation (2-3)
    24/3,
    13-15
    L51L3 MC (cont.), random number generation (cont.)
    v229/3, 8-10L51E1In class 2,4,7, Some notes on problem 7
    30/3, 13-15M33L4 Random number generation (cont.), variance reduction
    31/3,
    13-15
    Q21L5 Sequential Monte Carlo (SMC) methods
    v35/4, 8-10L51E2 In class 1,3,5(a-b)
    6/4, 10-12L51L6 SMC methods (cont.)
    7/4, 13-15L52L7 SMC methods (cont.)
    A key paper on SMC: Gordon et al. (1993)
    HA1 out, data needed.
    v419/4, 8-10L51E3
    20/4, 10-12Q31L8 Markov chain Monte Carlo (MCMC) (5.1,5.3,5.5-6)
    Two key papers on MCMC: Metropolis et al. (1953)
    Hastings (1970)
    21/4, 13-15L51L9 MCMC (5.2,5.7)
    v526/4, 8-10L51L10 MCMC (5.3-4)
    27/4, 10-12Q21E4
    28/4, 13-15L51L11 More on MCMC,
    stochastic modeling and Bayesian inference (6.1,10.1)
    HA1 in (28/4, 13:00:00)
    v63/5, 8-10L52E5
    4/5, 13-15Q21L12 MCMC for Bayesian computation (10.2-3,11)HA2 out

    Data and files for HA2 coal_mine.mat,
    atlantic.mat,
    est_gumbel.m
    9/5 Review HA1
    in
    5/5,
    13-15
    L51L13 Statistical models (6)
    v710/5, 8-10L51E6
    11/5, 13-15M33L14 Bootstrap (7)
    12/5, 13-15L51L15 Permutation tests (8), the EM algorithm
    v817/5, 8-10L51E7
    18/5, 13-15Q34L16 Advanced topicsHA2 in (18/5 13:00:00)
    19/5, 13-15L51No lecture Reserve, no lecture
    24/5 Review HA2 in (24/5, 12:00:00)
    30/5 Exam
  • Kursbeskrivning - course description
  • A list of theoretical questions for the exam
  • Exam, 9 June 2015 including solutions.
  • Exam, 17 August 2015 including solutions.
  • Exam, 1 June 2016 including solutions.
  • Exam, 30 May 2017 including solutions.
  • Lectures & Exercises: Johan Westerborn
  • E-postadress: johawes@kth.se
  • Examiner: Henrik Hult
  • E-mail: hult@kth.se

  • [Kursförteckning]     [Avdelningen Matematisk statistik]
    Sidansvarig: Jimmy Olsson
    Uppdaterad: 2015-06-09