Direkt zum Inhalt
Eine algebraische Fläche vom Grad 6 (eine "Sextik"), die 65 Singularitäten besitzt.


Sport Medicine

The analysis of multivariate time series about limp angles and muscle activity during Inline Speed-Skating was done using the with the Time Series Knowledge Mining. The resulting temporal patterns describe typical coordination patterns of muscles during repetitive movements. The pattern descriptions can be used for rehabilitation therapy and training optimization.


Time Series from  Inline Speed- Skating

Stock Analysis

The self-organization of 8.000 US stocks led to a new segmentation of the market. The knowledge gained from the groups was used to predict the probability of making profits. The forecasts were successful even when applied to currently generally falling stocks.


U-Matrix stocks

DNA Microarrays

In an analysis of DNA microarray we were able to extract the most relevant attributes from a set of 4000 genes. The reduction led to 19 genes particulariy well suited for prediction of an illness. This large decrease in complexity made it possible to perform a more detailed analysis of the function for those genes. First results indicate new insights into the growth of cancer for small kids.


U-Matrix Gene data

Customer Relationship Management

The clustering of customer behaviour applied to data from a mobile phone company produced a novel segmentation of the clients. The descriptions of the groups helped the company to identify the most interesting groups, e.g. all customers likely to stop the contract in the near future.


U-Matrix Mobile phones

Medical Diagnosis

An example from medicine is the diagnosis of lyme disease based on spinal fluid. The large red group on the picture represent healthy patient, the other groups turned out to be partly unknown forms of the infection.


U-Matrix Liquor Analysis

Zuletzt aktualisiert: 10.09.2007 · herrmanl

Fb. 12 - Mathematik und Informatik

Datenbionik (AG Ultsch), Hans-Meerwein-Straße 6, D-35032 Marburg
Tel. +49 6421/28-22185, Fax +49 6421/28-28902, E-Mail: databionics@informatik.uni-marburg.de

URL dieser Seite: http://www.uni-marburg.de/fb12/arbeitsgruppen/datenbionik/applications

Impressum | Datenschutz