Arbeitsgemeinschaft für Fuzzy-Logik und Soft Computing in Norddeutschland (AFN)

Otto-von-Guericke University of Magdeburg

Faculty of Computer Science
Institute of Knowledge Processing and Language Engineering
Neural Networks and Fuzzy-Systems

Universitaetsplatz 2
D-39106 Magdeburg

Contact: Prof. Dr. Rudolf Kruse

Phone: +49 391 67 18718

Fax: +49 391 67 12018

E-Mail: rudolf.kruse@cs.uni-magdeburg.de

URL: http://fuzzy.cs.uni-magdeburg.de/

Area of exhibition: Automation / Computer Science

Keywords

Fuzzy Logic, Neural Networks, Neuro-Fuzzy, Control Theory, Automatisation, Data Analysis, Data Mining, Fuzzy Clustering, Probabilistic Reasoning, Approximate Reasoning

Research Topics

Our group, led by Prof. Dr. Rudolf Kruse, examines phenomenas of uncertainty and vagueness in knowledge based systems, and develops and implements methods to cope with them. For this we do basic research on several methods like fuzzy systems, neural networks, probabilistic and approximate reasoning. These methods are especially suited to exploit the tolerance for uncertainty and vagueness in cognitive reasoning. By applying these methods it is possible to obtain easy to handle, robust, and low-priced solutions for certain areas of application.

Currently, our group is working on the following topics:

Neuro-Fuzzy Systems Fuzzy Data Analysis

Uncertainty and Vagueness in Knowledge Based Systems

More information is available from http://fuzzy.cs.uni-magdeburg.de/.

Our exhibits on the Forschungsforum '97

On the Forschungsforum '97 we present some software tools, which were developed by our research group. These tools implement the theoretical models which were evolved by our research activities.

FCLUSTER - A tool for fuzzy cluster analysis

FCLUSTER is a tool for fuzzy cluster analysis. Besides for fuzzy clustering it can be used to obtain a set of fuzzy rules which describe the underlying data. These rules can be used to build fuzzy systems like fuzzy classifiers or fuzzy controllers, for example.

NEFCLASS is available for UNIX.

NEFCLASS - Neuro-Fuzzy Classification

NEFCLASS is short for NEuro-Fuzzy CLASSification and it is used for data analysis by neuro-fuzzy models. It can learn fuzzy rules and fuzzy sets by supervised learning. NEFCLASS has the following features:

It represents a fuzzy classification system.
It learns fuzzy classification rules incrementally.
It learns fuzzy sets by using simple heuristics.
The learning algorithm does not afflict the semantics of the underlying fuzzy classifier. The system is always interpretable in terms of fuzzy classification rules.
NEFCLASS is available in two implementations (PC and UNIX).

NEFCON - Neuro-Fuzzy Control

NEFCON is a model for neuro-fuzzy control. It learns fuzzy rules and fuzzy sets by reinforcement learning. NEFCON has the following features:

It represents a normal fuzzy controller.
It is able to learn fuzzy rules incrementally or decrementally.
It learns fuzzy sets by using simple heuristics.
The learning alogrithm uses a linguistic error measure that is expressed by fuzzy rules.
The learning algorithm does not afflict the semantics of the underlying fuzzy controller. The system is always interpretable in terms of fuzzy rules.
NEFCON is available in three implementations (UNIX, Windows, MATLAB/SIMULINK).

Andreas Nürnberger (Andreas.Nuernberger@cs.uni-magdeburg.de), Aug. 11, 1997

Contact:	Prof. Dr. Rudolf Kruse
Phone:	+49 391 67 18718
Fax:	+49 391 67 12018
E-Mail:	rudolf.kruse@cs.uni-magdeburg.de
URL:	http://fuzzy.cs.uni-magdeburg.de/