TDT4173 Machine Learning and Case Based Reasoning

  • Main Page
  • Curriculum
  • Lectures
  • Previous Exams
  • Exercises

  • Exercise Information

    This page will be continuously be updated during the semester. New assignments are posted both here and at It's Learning.

    Exercises are administered by Lars Bungum

    There will be 4 exercises throughout the semester. You can qualify to enter the written exam without handing them in, but the results of the exercises count towards your final grade. All assignments are graded, and the value of each assignment is defined in the following way:

    • Assignment 1 is worth up to 3.33p.
    • Assignment 2 is worth up to 6.67p, as this is the "big" assignment.
    • Assignment 3 is worth up to 3.33p.
    • Assignment 4 is worth up to 3.33p.
    • Additionally, a student presentation of a paper will be seen as an extra source of 3.33 points.
    In total, the assignments can earn you a total of 20p. The written examination will be graded from 0p to 80p, and your total (between 0p and 100p) is the basis for the final grade. If your total is less than 40p you will fail the course, but if you have full score from the assignments, you will need only 20p (out of 80p) from the examination to pass.

    Individual work vs. working in groups

    All the exercises are individual tasks, so cooperation in groups is not approved there.

    Direct copying of any part of an exercise will result in immediately getting zero points for ALL the exercises of the course. This pertains to the copier as well as the source!


    Assistance hours: Wednesday 12:15-14:00, Room 359 IT-bygget

    Questions regarding exercises can also be emailed to Lars at any time. If you have problems with the exercise, make sure to ask for assistance. If you deliver an incomplete assignment without having asked for help, you will most likely fail.

    Delivery and approval

    Use It's Learning to deliver an exercise. Approval of the exercise is done by Teaching Assistants. Delivery deadlines will be published together with the corresponding exercises, but the assignments are typically due THURSDAYS BY 20:00 weak-and-a-half in advance. (See details below.)

    Deadlines are strictly enforced. This means that you will automatically get zero points for assignments handed in after the deadline.

    The deliveries should be submitted to the teaching assistant through it's learning. Submit the answers as a PDF file, and put any additional material (source code etc.) in a zip file. The names of the delivered files should be self-explanatory, e.g. report.pdf, code.java. Whenever you are asked to provide a piece of code in your delivery, take care of the following.

    • Use a programming language of your choice, unless other directions are given in the particular exercise.
    • Your code has to be runnable, whether as a standalone or a web-based application, or a plug-in or modification of an existing package. Please supply a brief (one or two sentences) description on how to run the code. In some exercises you may be asked to provide pseudocode, i.e. to show how you would solve the problem, without implementing it.
    • Please write the code so that it is easy to read and understand; add as many comments as necessary. You are criticized for the functionality and relevance of the code, and not for its beauty; however, the less readable the code, the more difficult it might be to check whether it works properly.
    Note that for an exercise to be approved, all parts of the exercise, i.e. all subexercises, must be answered adequately.

    Exercise Release dateDue dateSolution
  • Exercise 1 - Machine Learning fundamentals
  • file
  • Exercise 2 - THE BIG EXERCISE: Regression
  • file
  • Exercise 3 - Bayesian Learning
  • file
  • Exercise 4 - Case based reasoning
  • file

    Editor: Head of department : Maria Letizia Jaccheri   Contact address: Helge Langseth   Page updated: