Analyzing Infant Cries using a Committee of Neural Networks in Order to Detect Hypoxia Related Disorder

G.D. Magoulas (Editor), M. Poel, G. Dounias (Editor), T. Ekkel, D.A. Linkens (Editor)

  • 3 Citations

Abstract

Based on the hypothesis that the sound of the infant cry contains information on the infant's health status, research is done to improve classi¯cation of neonate crying sounds into categories called 'normal' and 'abnormal' - the latter referring to some hypoxia- related disorder. Research in this field is hindered by lack of test cases and limited understanding of feature relevance. The research described here combines various ways of dealing with the small data set problem. First, feature pre-selection is done using sequential backwards elimination of possible combinations where the performance of the set of features is tested by a Probabilistic Neural Network which has the advantage of fast learning. These features are then fed into a neural network committee consisting of Radial Basis Function Neural Networks. Bootstrapping is used to generate the committee. This construction yields a multi-classifier system with an overall classi¯cation performance of 85% on the data set, an increase of 34% with respect to the a priori probability of 51%. Several leave-1-out experiments for Linear Discriminant Analysis (LDA), Support Vector Machines (SVM) and Neural Networks (NN) have been conducted in order to compare the performance of the multi-classifier system. Keywords: Infant Cry Analysis; Feature reduction, Neural Networks; Support Vector Machines
Original languageEnglish
Pages (from-to)397-410
Number of pages14
JournalInternational journal on artificial intelligence tools
Volume15
Issue number3
DOIs
StatePublished - Jun 2006

Fingerprint

Neural networks
Support vector machines
Classifiers
Positive ions
Discriminant analysis
Health
Experiments

Keywords

  • METIS-238672
  • IR-63414
  • EWI-6855

Cite this

Magoulas, G.D. (Editor); Poel, M.; Dounias, G. (Editor); Ekkel, T.; Linkens, D.A. (Editor) / Analyzing Infant Cries using a Committee of Neural Networks in Order to Detect Hypoxia Related Disorder.

In: International journal on artificial intelligence tools, Vol. 15, No. 3, 06.2006, p. 397-410.

Research output: Scientific - peer-reviewArticle

@article{2c57905ebf6340138e92cb3b270dd840,
title = "Analyzing Infant Cries using a Committee of Neural Networks in Order to Detect Hypoxia Related Disorder",
abstract = "Based on the hypothesis that the sound of the infant cry contains information on the infant's health status, research is done to improve classi¯cation of neonate crying sounds into categories called 'normal' and 'abnormal' - the latter referring to some hypoxia- related disorder. Research in this field is hindered by lack of test cases and limited understanding of feature relevance. The research described here combines various ways of dealing with the small data set problem. First, feature pre-selection is done using sequential backwards elimination of possible combinations where the performance of the set of features is tested by a Probabilistic Neural Network which has the advantage of fast learning. These features are then fed into a neural network committee consisting of Radial Basis Function Neural Networks. Bootstrapping is used to generate the committee. This construction yields a multi-classifier system with an overall classi¯cation performance of 85% on the data set, an increase of 34% with respect to the a priori probability of 51%. Several leave-1-out experiments for Linear Discriminant Analysis (LDA), Support Vector Machines (SVM) and Neural Networks (NN) have been conducted in order to compare the performance of the multi-classifier system. Keywords: Infant Cry Analysis; Feature reduction, Neural Networks; Support Vector Machines",
keywords = "METIS-238672, IR-63414, EWI-6855",
author = "G.D. Magoulas and M. Poel and G. Dounias and T. Ekkel and D.A. Linkens",
year = "2006",
month = "6",
doi = "10.1142/S0218213006002734",
volume = "15",
pages = "397--410",
journal = "International journal on artificial intelligence tools",
issn = "0218-2130",
publisher = "World Scientific Publishing Co. Pte Ltd",
number = "3",

}

Analyzing Infant Cries using a Committee of Neural Networks in Order to Detect Hypoxia Related Disorder. / Magoulas, G.D. (Editor); Poel, M.; Dounias, G. (Editor); Ekkel, T.; Linkens, D.A. (Editor).

In: International journal on artificial intelligence tools, Vol. 15, No. 3, 06.2006, p. 397-410.

Research output: Scientific - peer-reviewArticle

TY - JOUR

T1 - Analyzing Infant Cries using a Committee of Neural Networks in Order to Detect Hypoxia Related Disorder

AU - Poel,M.

AU - Ekkel,T.

A2 - Magoulas,G.D.

A2 - Dounias,G.

A2 - Linkens,D.A.

PY - 2006/6

Y1 - 2006/6

N2 - Based on the hypothesis that the sound of the infant cry contains information on the infant's health status, research is done to improve classi¯cation of neonate crying sounds into categories called 'normal' and 'abnormal' - the latter referring to some hypoxia- related disorder. Research in this field is hindered by lack of test cases and limited understanding of feature relevance. The research described here combines various ways of dealing with the small data set problem. First, feature pre-selection is done using sequential backwards elimination of possible combinations where the performance of the set of features is tested by a Probabilistic Neural Network which has the advantage of fast learning. These features are then fed into a neural network committee consisting of Radial Basis Function Neural Networks. Bootstrapping is used to generate the committee. This construction yields a multi-classifier system with an overall classi¯cation performance of 85% on the data set, an increase of 34% with respect to the a priori probability of 51%. Several leave-1-out experiments for Linear Discriminant Analysis (LDA), Support Vector Machines (SVM) and Neural Networks (NN) have been conducted in order to compare the performance of the multi-classifier system. Keywords: Infant Cry Analysis; Feature reduction, Neural Networks; Support Vector Machines

AB - Based on the hypothesis that the sound of the infant cry contains information on the infant's health status, research is done to improve classi¯cation of neonate crying sounds into categories called 'normal' and 'abnormal' - the latter referring to some hypoxia- related disorder. Research in this field is hindered by lack of test cases and limited understanding of feature relevance. The research described here combines various ways of dealing with the small data set problem. First, feature pre-selection is done using sequential backwards elimination of possible combinations where the performance of the set of features is tested by a Probabilistic Neural Network which has the advantage of fast learning. These features are then fed into a neural network committee consisting of Radial Basis Function Neural Networks. Bootstrapping is used to generate the committee. This construction yields a multi-classifier system with an overall classi¯cation performance of 85% on the data set, an increase of 34% with respect to the a priori probability of 51%. Several leave-1-out experiments for Linear Discriminant Analysis (LDA), Support Vector Machines (SVM) and Neural Networks (NN) have been conducted in order to compare the performance of the multi-classifier system. Keywords: Infant Cry Analysis; Feature reduction, Neural Networks; Support Vector Machines

KW - METIS-238672

KW - IR-63414

KW - EWI-6855

U2 - 10.1142/S0218213006002734

DO - 10.1142/S0218213006002734

M3 - Article

VL - 15

SP - 397

EP - 410

JO - International journal on artificial intelligence tools

T2 - International journal on artificial intelligence tools

JF - International journal on artificial intelligence tools

SN - 0218-2130

IS - 3

ER -