Uni-Logo
Sie sind hier: Startseite Daten Lehrstühle CGNM Machine Learning on the Grid, especially Classification of Complex Data
Artikelaktionen

Machine Learning on the Grid, especially Classification of Complex Data

The Computer-based New Media Group works on various problems from machine learning and data mining, especially on classification problems for complex data with high demands on resources, usually cpu time, but for some problems also memory.

For example, for our main application, recommender systems in e-commerce, i.e., methods for automatic personalization and selling, multi-class models for a very high number of alternatives (2000 and more) have to be learned. Among the best performing models are ensemble models consisting of millions of component models.

CGNM is a member of the Black Forest Grid Initiative that aims at developing a grid infrastructure in a bottom-up style by joining efforts of several groups.

Beneath these core applications, we use the Grid infrastructure for a broad set of applications in web mining, e-learning, medical informatics, data mining in engineering, etc.

Additionally to classification methods we research and implement Bayesian network models, mixture models, cluster analysis and pattern mining, especially high-performance algorithms and algorithms for complex pattern spaces.

Benutzerspezifische Werkzeuge