Theses & Internships
This page is for students seeking a project, a seminar subject or a thesis on topics of the KMD group. The page is on English, but topics can also be written on German.
Auf diese Webseite können Studierende Themen und Kontaktpersonen für Projekte, Seminare und Abschlussarbeiten bei der KMD Arbeitsgruppe finden. Alle Themen können auch auf Deutsch geschrieben und betreut werden.
You look for a team project or IT software project, for a seminar work, for a subject that is appropriate for a bachelor or master thesis, or for an individual project? In this page, you find the topic areas we offer for our students. These topics are closely coupled with our research; some of them are in cooperation with companies, medical institutions or other external partners. For courses relevant to these topics, please have a look here.
In medical data mining, we use learning algorithms to solve concrete problems of disease diagnosis, impact of medical treatment or identification of risk factors for a disorder. As explained in our research page on Medical Mining, we study epidemiological data in cooperation with external partners, mainly medical research institutes. We offer topics for seminars, theses and projects on the discovery of subpopulations that share the same risk factors or respond similarly to a treatment. Our topics encompass literature surveys, evaluation of existing methods and development of new ones.
For more information on this thematic area, please contact Tommy Hielscher.
Recommenders are intelligent systems that use data mining methods to learn users' preferences e.g. in online shops, as it known from Amazon, YouTube, Netflix, etc. They build profiles of users based on user past behaviour, and they use those profiles to predict users' future preferences. These predictions are then exploited to create a personalized set of recommendations for a given user. The recommended items may be movies, literature, web pages, friends, products, advertisements and many more. Nowadays, when information overload increasingly becomes a problem for all users, recommender systems provide a necessary solution by reducing the number of books, movies etc. to a small relevant selection that a user can handle.
For more information on this thematic area, please contact Pawel Matuszyk.
This list of thematic areas is not complete. Topics on active learning and drift mining will be added (contact: Georg Krempl). For topics on stream mining algorithms (classification, clustering, semi-supervised learning on news, opinionated documents and on conventional recordings) please contact Myra Spiliopoulou.
You have a topic yourself?
Inspire us! Please contact Myra Spiliopoulou.
You have some coarse idea, but are not sure?
We can help you! All KMD members can give you more information on their topic areas, and tell you what background you need for them. If you need advice on another subject area, please contact Myra Spiliopoulou.