Selection of an artificial neural network model to diagnosis mouth-breathing children

Selection of an artificial neural network model to diagnosis mouth-breathing children

Author Mancini, Felipe Autor UNIFESP Google Scholar
Pisa, Ivan Torres Autor UNIFESP Google Scholar
Yi, Liu Chiao Autor UNIFESP Google Scholar
Pignatari, Shirley Shizue Nagata Autor UNIFESP Google Scholar
Azevedo, L. Google Scholar
Londral, A. R. Google Scholar
Institution Universidade Federal de São Paulo (UNIFESP)
Abstract A number of factors can lead to changes in body posture, basically determined by alterations in the natural curvature of the spine. Such changes, in turn, may also result in secondary health problems. Mouth breathing is thought to be one of these problems. Experiments with healthy nasal breathing individuals have showed that when they are forced to breathe through their mouth only the natural shape of their spine curves change. However the characterization of the spine curvature in mouth breathers has not been done yet and the matter lies on the personal experience of the health professional. This study reports on the preliminary findings of a broader research which attempts to characterize the changes in the behaviour of the spine, caused by mouth breathing, by using artificial neural network modelling and data from 52 subjects. Four different models - backprogation, learning vector quantization (LVQ), and self-organizing map (SOM) - were tested for best performances in sensitivity and specificity in diagnosing mouth and nasal breathing children. Competitive-leaming-based algorithms - LVQ and SOM - presented the best performance for current data set.
Keywords Artificial neural networks
Mouth breathing
Clinical decision support systems
Language English
Date 2008-01-01
Published in Healthinf 2008: Proceedings Of The First International Conference On Health Informatics, Vol 2. Setubal: Insticc-inst Syst Technologies Information Control & Communication, p. 197-200, 2008.
Publisher Insticc-inst Syst Technologies Information Control & Communication
Extent 197-200
Access rights Closed access
Type Conference paper
Web of Science ID WOS:000256698200035

Show full item record


File Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)




My Account