Democritus University of Thrace (DUTH)
Ανάπτυξη ευφυών αλγόριθμων στην ιατρική διαγνωστική
Abstract
dc:descriptionIn the last three decades, there is a vivid interest of many researchers about algorithms with natural procedure similarities. Artificial Neural Networks, Genetic Algorithms and Evolutionary Programming are the wide known algorithms, that have been developed in order to solve complex combinational problems. Although the aforementioned algorithms are techniques that used in Artificial Intelligence recently, are also referred in bibliography as Computational Intelligence algorithms and are used in varied scientific areas. One of these areas is the Medicine. Nowadays it is attempted strenuous and specific efforts about the development of methods, which would be adequate in order to diagnose an illness based on clinical symptoms and laboratorial examinations. The thesis studies the implementation of computational intelligent methods to medical diagnosis, that would comprise a helpful tool for physicians. In additional, pruning techniques are used for elimination of diagnostic factors that take into account for assessment of a case, reducing the redundancy information of medical data. The implemented computational algorithms are based on artificial neural networks, genetic algorithms and their combination. The aforementioned methods are applied to: • pediatric urology, and specific urinary tract infection and vesicoureteral reflux, • abdominal pain in childhood and, • osteoporosis. The proposed algorithms are the first efforts for application of intelligent method to specific diseases. Diversity of ANN architectures are implemented assessed for correct prediction of outcome for each of aforementioned diseases, based on available cases during the study. The ANN architecture with the best prognostic performance for each illness, is used by the physicians in order to ease their clinical examination. The obtained results proves that both ANNs and GAs have the ability to be incorporated in varied fields of Medicine. This fact does not abolish the human role in clinical practice. Medicine is a science that the human factor has a preeminent role. The continual technological growth facilitates the physician’s work and contributes to accurate and timely cure of diseases. The human brain will have the master role, as it would not be ignored that the machines and computers are human contrivances.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Mantzaris, Dimitrios
- Μαντζάρης, Δημήτριος
Subjects
dc:subject × 18- Τεχνητή νοημοσύνη
- Τεχνητά νευρωνικά δίκτυα
- Γενετικοί αλγόριθμοι
- Πρόβλεψη κυστεοουρητηρικής παλινδρόμησης
- Εκτίμηση οξείας σκωληκοειδίτιδας
- Πρόβλεψη κινδύνου οστεοπόρωσης
- Εξελικτικά νευρωνικά δίκτυα
- Artificial intelligence
- Artificial neural networks
- Genetic algorithms
- Vesicoureteral reflux prediction
- Acute appendicitis estimation
- Osteoporosis risk factor prediction
- Evolutionary neural networks
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/26005