|
Title:
|
CLINICAL DECISION SUPPORT SYSTEM FOR CARDIAC VIABILITY DIAGNOSIS USING NEURAL NETWORKS |
|
Author(s):
|
Javier Moya , Pau Micó , Vicent Bodí |
|
ISBN:
|
978-972-8924-97-3 |
|
Editors:
|
Hans Weghorn and Pedro Isaías |
|
Year:
|
2009 |
|
Edition:
|
V I, 2 |
|
Keywords:
|
Cardiac viability diagnosis, neural network. |
|
Type:
|
Full Paper |
|
First Page:
|
59 |
|
Last Page:
|
66 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The present paper demonstrates how multilayer perceptron neural networks can be useful for studying the cardiac
viability, and therefore for designing clinical diagnose support systems. Cardio-magnetic resonances were performed to a
number of patients in order to obtain six indexes and a viability target. Such data was then cleaned up and validated as a
corpus. Four different network models were then proposed, aiming to establish their performance; each network needed,
however, the restructuring of the corpus data, and, consequently, different validation procedures. Their performance was
measured with the help of two different error rates: a global classification error and a viable class error. Four learning
algorithms were put to the test on the simplest of the network models, of which the vanilla backpropagation one proved
to give best results. Every contextual model was tested for every AHA segment, and significantly lower error rates were
found for them, proving that multilayer perceptron neural networks do have a role in clinical diagnosis support systems
for viability diagnosis. |
|
|
|
|
|
|
