Bogdan's Research:
Towards the Construction of Fuzzy Information Spaces
Problem Statement
During the fall of 1992, I was given the following problem:
Given the electroencephalographic (EEG)
signals recorded during sleep,
how do we characterize and quantify transitions between sleep stages?
In a matter of few weeks I've been supplied with the megabytes of binary files
containing digitized EEG signals and a few basic texts on sleep physiology.
About ten months later I realized that the initial problem may have a solution
if one could answer another, more general question:
How do we characterize and quantify complex
physical processes from their signals
keeping both qualitative and temporal properties of dynamics within the
scope of analysis?
This text provides a possible answer to this question.
The rest of this page is still
under ! Come again soon.
First Steps
EEG signals
Time-frequency images of sleep
Dynamic Systems and their Processes
Temporal Fuzzy Sets
Hidden Process Model
Fuzzy Information Space
Where and How to Apply these Concepts?
Back to Analysis of Sleep Dynamics
Feature space construction
Simulations
A real-world example
Current and Future Work
Bogdan R. Kosanovic, Dept. of Electrical Eng.,
University of Pittsburgh, bogdan@neuronet.pitt.edu