ELECTRAS


Background

Aim

In science huge amounts of data are produced. The intrinsic information of these data is often difficult to grasp. In many cases, not only the data themselves are interesting, but especially the intrinsic relationships among the data. Thus, methods have to be employed which enable a deeper analysis of huge data sets. Methods used for the analysis of data are statistical methods and non-linear methods, such as neural networks.

This web-based application offers an interface to various kinds of data analyses. Besides the application of different statistical methods, e.g. linear regression, correlation, the data can be analyzed through several kinds of neural networks (e.g. Kohonen Network, Counterpropagation Network, Backpropagation Network). The data can also be weighted and scaled and transformed by different transformations, e.g. Fourier Transform, Wavelet Transform.
The results of the analyses are visualized e.g through VRML graphics and can also be listed in a report or given in an output file.

 

 

Last modified Dec. 10 th, 2001