Aristotle University Of Thessaloniki (AUTH)
ΠΡΟΣΟΜΟΙΩΣΗ ΕΓΚΕΦΑΛΙΚΗΣ ΔΡΑΣΤΗΡΙΟΤΗΤΑΣ ΜΕ ΧΡΗΣΗ ΗΛΕΚΤΡΟΝΙΚΟΥ ΥΠΟΛΟΓΙΣΤΗ
Abstract
dc:descriptionIN THIS STUDY WE INTRODUCE SOME ARTIFICIAL NEURAL NETWORKS (ANNS) LIKE BP, LVQ,CP AND SOM WHICH HAVE SIGNIFICANT ABILITIES IN GENERALIZING ON UNKNOWN VECTORS. FOR THE PURPOSE OF OUR STUDY WE FORMULATED INDEPENDENT EEG SETS OF DATA FOR TRAINING AND TESTING ANNS THAT HAD SIMILAR PROPERTIES. EACH SET INCLUDED BOTH NORMAL AND EPILEPTIC DATA. SOME OF THE ANNS WERE TRAINED AND TESTED USING THE CORRESPONDING POWER SPECTRUM DENSITY (PSD). RESULTS SHOW THAT ANNS CAN BE USED IN RECOGNIZING AND CLASSIFYING EPILEPTIC LONG TERM EEG WITH SUFFICIENT ACCURACY. (80%-90% OF THE TOTAL TESTING SET). A SYSTEM BASED ON ANNS COULD BE VERY HELPFULIN THE TASK OF RECOGNIZING EPILEPTIC WAVEFORMS.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 1989
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ρίζος, Γεώργιος
Subjects
dc:subject × 14- Networks, Artificial neural
- AUTOMATED EEG ANALYSIS
- PATTERN RECOGNITIONAND CLASSIFICATION
- ΑΝΑΓΝΩΡΙΣΗ ΚΑΙ ΤΑΞΙΝΟΜΗΣΗΠΡΟΤΥΠΩΝ
- ΑΥΤΟΜΑΤΟΠΟΙΗΜΕΝΗ ΑΝΑΛΥΣΗ ΗΕΓ
- Δίκτυα, Τεχνητά νευρωνικά
- Ιατρική και Επιστήμες Υγείας
- Βασική Ιατρική
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Medical and Health Sciences
- Basic Medicine
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/2431