|Website:||website containing additional information|
|Period:||period 4 (week 17 through 26, i.e., 23-4-2018 through 29-6-2018; retake week 28)
|Participants:||up till now 46 subscriptions|
|Schedule:||Official schedule representation can be found in Osiris|
|Contents:||Pattern mining is characteristic for data mining. Whereas data analysis is usually concerned with models – i.e., succinct descriptions of all data – pattern mining is about local phenomena. Patterns describe – or even are – subgroups of the data that for some reason are deemed interesting; a description and a reason that usually involves some – if any -- of the variables (attributes features) rather than all.
In the past few decades – the total existence of data mining – pattern mining has proven to be a fruitful research area with many thousands of papers describing a wide variety of pattern languages, interestingness functions, and even more algorithms to discover them. However, there is a problem with pattern mining. Databases tend to exhibit many, very many patterns. It is not uncommon that one discovers more patterns than one has data. Hardly an ideal situation. Hence, the rise of pattern set mining. Can we define and find relatively small, good sets of patterns?
In this course we’ll start with a brief discussion of pattern mining. After that we discuss parts of the literature on pattern set mining; only parts because there is too much to discuss it all. What types of solutions have been proposed? How do they work and, actually, do the work?
|Literature:||In each class we will discuss one or more papers from the pattern and pattern set mining literature. Links to the articles will be provided on the website of the course and will be accessible from within the university's network. |
|Course form:||The term ``discussion’’ above should be taken literally. Rather than using class time for lectures taught by me or by one of the students, we will meet to discuss. Appointed papers should be read by each and everyone of us before class and during class we discuss those papers. What problem did the authors aim to solve, how did they approach it and did they succeed? It is not so much about my answers to these questions, it is about your answers
|Exam form:||The goal of this course is that you understand what pattern set mining is all about. The best way to prove that you do is by writing an essay on it. So, next to actively participating in the meetings that is what you have to do.|
|Minimum effort to qualify for 2nd chance exam:||Om aan de aanvullende toets te mogen meedoen moet de oorspronkelijke uitslag minstens 4 zijn.|