MTAT.03.251 Graph Mining
MTAT.03.251 Andmekaeve graafides

  • Seminars: Wednesdays K 12.15 Liivi 2-122
  • Office hours: Mondays Liivi 2-326. Please arrange exact details of the meeting by email.
  • Questions:
  • The first seminar starts on the 10th of February
  • The course is based on the book Eric D. Kolaczyk. Statistical Analysis of Network Data.
  • Lectures and seminar sessions are held in English or in Estonian depending on circumstances.
  • Final reports should be written preferably in English.

What and Why

The course aims to give a overview of basic graph mining techniques starting form the graph construction and ending with prediction of graph structures. Graph mining is relatively new research area. Although graphs are natural way to represent connections and associations between objects, they are also complex discrete objects compared to simple arrays and matrices. As a result, standard statistical techniques and machine learning algorithms have to be modified to run on top of graphs. In this course, we view basic ways how to overcome this by using graph specific methods and various ways to flatten the complex structure by feature extraction. Another reason why graph mining is relatively new is purely pragmatical. Although large network structures have been present for a long time, only recent changes in technology have made it possible to collect such kind of data. For example, only recent network-oriented applications such as instant messengers, social network websites and geo-tagging, have opened an opportunity to observe and study large-scale social interactions. Also, lack of efficient algorithms and lack of raw computing power have been hindering factors. As such the topic is highly suitable for students from computer science and statistics. In particular, note that elements of graph mining are heavily used in system biology and bioinformatics and several major projects done in cooperation between department of computer science and competence center STACC are closely related to study of social networks.


There are no formal prerequisites to the seminar. However, basic knowledge in programming, basic math and graph theory is advisable. Also, one needs reasonable English skills to complete the course report. If the formal requirements of the ÕIS do not permit registration then write me an email or talk with me. After that we decide whether to enroll you or not.

To pass the course

  • You have to give at least one presentation about the topics covered in this course.
  • You must write an illustrative demo application for covering the techniques you present. You do not have to implement the algorithms by yourself. It is sufficient if you use packages of R or some other computing software.
  • You must actively participate in most seminars or otherwise you do not pass the seminar. Namely, student gets grade F if he or she misses 3 or more seminars. In reasonable circumstances, it is possible to compensate missed seminars by extra work. Details are determined by individual agreements with the lecturer.
  • You must do a project work. The project work will be graded in the scale from A to F and it will form the basic grade for the course.
  • The requirements for the project work are standard: they must cover the main research problem and methods so that they would be understandable for the fellow students. The description of experiments should be detailed enough so that it is repeatable by others and your own contribution is clearly visible. Results should be presented together with clear interpretation.
Edit: header| contents| footer| sidebar