Hide menu

732A89 Computational Statistics

Course information


This is the coursepage for the course in Computational Statistics.

The first course occasion was a lecture in room U6 on Wednesday 2025-01-22 13:15-15:00.

This course is an obligatory course in the master's program Statistics and Machine Learning. Other students can apply for it as a single-subject course.

The target student group are those who are familiar with statistics, have good mathematical background and practical experience of programming in some computer language.

The programming language of the course is R.

The course literature is:

Course structure


The course contains 8 lectures, 6 computer labs, and 3 seminars. Attendance at the seminars is mandatory.

Lectures, computer labs, and seminars will be on campus; see TimeEdit for dates, times, and locations.

Six topics (lecture, lab, and seminar hour) will focus on different aspects of computational statistics and students will then work in groups of two and two. Two lectures (LM1-LM2) will focus on a recall of basic mathematics required for the course.

The best thing is that the students work with their own computer.

The course contains three teaching activities:

  • Lecture (Fö) Introduction of new concepts.

  • The following content will be presented in the lectures.

    Topic 1: Unidimensional optimization and computer arithmetics

    Read: GH, Ch. 2.1; Gen, Ch. 2, 6.1
    Single subject course students: Self study of R syntax and basic functions, read and test the code from here


    Topic 2: Multidimensional optimization

    Read: GH, Chapter 2.2; Gen, Chapter 6.1 and 6.2


    Topic 3: Random number generation

    Read: GH, Chapter 6; Gen, Chapter 7.1-7.3


    Topic 4: Monte Carlo methods, MCMC

    Read: GH, Chapter 7; Gen, Chapter 7.3-7.4


    Topic 5: Bootstrap and hypothesis testing

    Read: GH, Chapter 9.1-9.3, 9.8; Gen, Chapter 12, 13


    Topic 6: EM algorithm and stochastic optimization

    Read: GH, Chapter 3.3-3.4, 4.1-4.2.2, Gen, Chapter 6.3, 6.6, 14.3

    Lecture Date and time Material
    L1 2025-01-22 13:15 Lecture 1
    R code for Newton
    Bisection video
    LM1 2025-01-27 08:15 Analytic optimization
    Basic matrix algebra
    Determinants
    Some R code connected to this lecture
    L2 2025-01-28 10:15 Lecture 2
    R code for steepest ascent example
    LM2 2025-02-03 08:15 Lecture slides
    L3 2025-02-04 10:15 Lecture 3
    L4 2025-02-11 10:15 Lecture 4
    R code for Metropolis algorithm (Example 1 from lecture)
    R code for ellipse from Gibbs example
    L5 2025-02-18 10:15 Lecture 5
    R code analysis of July-rain
    R-datafile julyrain.RData
    Datafile kresseertrag.dat
    L6 2025-02-26 10:15 Lecture 6
    R code for criterion in Lab 6 Q2
    R code for EM-algorithm of Lecture 6 (use it in Lab 6 Q1)
    R-datafile bankdata.RData
    R-datafile threepops.RData


  • Computer lab (DATALAB) Individual computer lab with individual help.
  • Attendance of the lab sessions is not obligatory but it might be difficult to complete the labs without supervision so it is recommended to attend these sessions.
    Students must discuss their lab solutions in a group (groups will be setup on the first Lab) and compile a collaborative report showing the results and the code. If the group is presenting then it should be done in such a way that the report document can directly be used for a presentation at the seminar (but code is still to be provided). Attention: there is a deadline for such a report! The document should clearly state the names of the students that participated in its compilation. This report should be submitted via LISAM as a .PDF (alternatively in case of problems emailed to one of the responsible staff) before the report deadline.
    ALL will be CHECKED for plagarism (also with respect to past labs)!
    The file should be named Group X.pdf where X is the group number. Please also include your names in the report.
    The collaborative reports are corrected and graded by the teacher. A student is PASSED on the lab if the group report is PASSED.

    All group members have to contribute to, understand and be able to explain all aspects of the work. In case some member(s) of a group do not contribute equally this has to be reported and in this situation a formal group work contract will be signed, stipulating the consequences for further unequal contributions.

    If you miss the deadline for a lab solution, you must submit the solution anyway, and in this case an additional assignment will also be given. There is a second deadline of 23:59 08 April 2025 for submitting corrections for all the hand-ins. There is a final deadline of 23:59 29 April 2025 for all the hand-ins. After this date NO submissions nor corrections will be accepted.

    Students who have participated in the last course (fall 2023) and need to submit some lab can write me (frank.miller at liu) an email. I will add you then to this year's LISAM and assign a group number to you such that you can submit your lab report.

    Assignment no. Instructions Lecture Lab Deadline Seminar Presenting groups Opposing groups
    1 Lab1.pdf 2025-01-22 2025-01-24 2025-01-29 2025-02-12 13
    12
    7
    9
    2 Lab2.pdf 2025-01-28 2025-01-29/31 2025-02-04 2025-02-12 4
    10
    1
    5
    11
    14
    3 Lab3.pdf 2025-02-04 2025-02-07 2025-02-11 2025-02-26 7
    6
    3
    8
    10
    1
    4 Lab4.pdf 2025-02-11 2025-02-14 2025-02-18 2025-02-26 2
    5
    4
    13
    5 Lab5.pdf 2025-02-18 2025-02-21 2025-02-25 2025-03-11 11
    9
    6
    2
    6 Lab6.pdf 2025-02-26 2025-03-03 2025-03-05 2025-03-11 14
    8
    3
    12

  • Seminar (SE) Each student group will present/oppose their labs and we will discuss eventual problems.

  • Attendance of the seminars is mandatory; each seminar involves a discussion of the latest labs.
    Each group will be assigned to present their lab report and to be opponents at some seminar; see the table above for the group numbers (but note that slight changes might be necessary for the later seminars). The opposing group is assigned to the group report which is in the same line in the table above.
    The presenting group is expected to present their solution using around 5-6 minutes per question (i.e., 10-12 minutes in total) and the two members should both present some part. They should show slides or a document, either using an own laptop or sending the slides in advance to the examiner by email (in the latter case, in .PDF or .ppt format).
    Each member of the opponent group should prepare at least two questions, comments, improvements, or suggestions to the group report they are assigned to. It is important that the opponents submit their own group report for the same lab within the deadlines as they will be provided with the group report of another group prior to the seminar.
    An inappropriately done presentation or opposition can result in an additional assignment or even lab failure for the appropriate group!
    Please contact the examiner in advance if you are speaker or opponent and cannot come to the seminar you are responsible for.
    If you miss a seminar, you need to submit a completely correct solution of the lab(s) discussed at this seminar, and an additional assignment will be given.
    If you miss two or more seminars, the laboratory part is graded FAILED and you are forwarded to attend the same series of the seminars in the next year course.

Software


The students are suggested to work using their own computers. For this course the following software is needed. Everything is open source and free.

Information on how to install R and R-Studio: - Windows, - Mac - Linux/Ubuntu

Computer exam


The exam will be a computer exam on 2025-03-24, from 8:00 to 13:00. Re-exams will be on 2025-05-16 and on 2024-08-29, between 8:00 and 13:00 in both cases.

Exam is akin to labs, the only difference is that you are given certain amount of hours and that you are not allowed to communicate with others in any way. The material that you are allowed to use at the exam: course books and a handwritten one-page A4 document (front page, only) with own notes. PDFs of the course book Givens and Hoeting (2013) will be available during the exam for the chapters 1-4, 6, 7, 9, index, frontmatter (incl. table of contents). PDFs of the lecture slides for L1, L2, L3, L4, L5, and L6 will be available during the exam as well. Further, an R reference card will be available.

Previous exams can be found here.

Staff


  • Martin Andrae, Bayu Brahmantio, teaching assistants
  • Frank Miller, examiner and lecturer

Page responsible: Frank Miller
Last updated: 2025-03-18