Aktuelle Seite:
Seminar: Mining Metadata - Details

  • Detaillierte Informationen über die Veranstaltung werden angezeigt, wie z.B. die Veranstaltungsnummer, Zuordnungen, DozentInnen, TutorInnen etc. In den Detail-Informationen ist unter Aktionen das Eintragen in eine Veranstaltung möglich.

  • link-extern Weiterführende Hilfe
Sie sind nicht angemeldet.

Mining Metadata

Allgemeine Informationen

Veranstaltungsnummer 8.3160
Semester SS 2014
Aktuelle Anzahl der Teilnehmenden 19
Heimat-Einrichtung LE Cognitive Science
Veranstaltungstyp Seminar in der Kategorie Offizielle Lehrveranstaltungen
Erster Termin Di , 22.04.2014 12:00 - 14:00, Ort: 35/E21
Sprache Englisch
Literatur http://hcil2.cs.umd.edu/newvarepository/
Kommentar /Beschreibung (englisch) The course is structured as follows:
First, we will get an overview of current data mining techniques. These techniques will be presented by the participants in "short talks", i.e., 10 -15 min presentations.
Afterwards, participants will work in groups and apply data mining techniques to "real" data sets.
Each group will present their results in a mid-term presentation and a final presentation. Additionally, each group writes a short report about their results and the work distribution between the individual participants.
Voraussetzungen (englisch) Participants should have programming skills since the course will include a practical component
Contact Hours 2
Sonstiges Grades will be calculated as
Short talk: 15%
Mid-term presentation: 15%
Final presentation: 25%
Results: 25%
Report: 5%
Work performance: 15%
ECTS-Punkte 4




Dienstag: 12:00 - 14:00, wöchentlich (ab 22.04.2014)





Prerequisites: Basic knowledge of Machine Learning

Data mining based on metadata seems a more complex task than applying data mining techniques on information which conveys semantics and context. However, for example, social networks, can be analyzed from their structure alone, without actual access to the data transferred within a social circle. This seminar will deal with different approaches for mining metadata to gain insight. Topics include approaches in the areas of image browsing, social networks, and problems related to the topic of "big data".