Auditory brainstem evoked potential classification for threshold detection by neural networks. II. Effects of input coding, training set size and composition and network size on performance

D. Alpsan, Ozcan Ozdamar

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

A feedforward neural network with one hidden layer is applied to the problem of brainstem auditory evoked potential classification. Network performances were tested separately both on subject-dependent samples (drawn from the same subjects from which the training set was derived) and on subject-independent samples (drawn from subjects from which no data was included in the training set), and compared. The results indicate that increasing the training set size improves performance, and human selected training sets give better results than randomly selected sets. Different encoding schemes of signal representation yielded a wide range of correct recognition rates. A spectro-temporal representation was very successful in terms of generalization to novel input and for producing high recognition rates for both classification categories. Although the networks were overly complex and trained in the memorization mode, they displayed feature extracting and generalization capabilities.

Original languageEnglish
Pages (from-to)83-93
Number of pages11
JournalAutomedica
Volume15
Issue number1
StatePublished - Dec 1 1992
Externally publishedYes

Fingerprint

Brain Stem Auditory Evoked Potentials
Bioelectric potentials
Neural networks
Feedforward neural networks
Network performance
Chemical analysis

ASJC Scopus subject areas

  • Biophysics
  • Bioengineering

Cite this

@article{6cc185ec338746359568da0861e90511,
title = "Auditory brainstem evoked potential classification for threshold detection by neural networks. II. Effects of input coding, training set size and composition and network size on performance",
abstract = "A feedforward neural network with one hidden layer is applied to the problem of brainstem auditory evoked potential classification. Network performances were tested separately both on subject-dependent samples (drawn from the same subjects from which the training set was derived) and on subject-independent samples (drawn from subjects from which no data was included in the training set), and compared. The results indicate that increasing the training set size improves performance, and human selected training sets give better results than randomly selected sets. Different encoding schemes of signal representation yielded a wide range of correct recognition rates. A spectro-temporal representation was very successful in terms of generalization to novel input and for producing high recognition rates for both classification categories. Although the networks were overly complex and trained in the memorization mode, they displayed feature extracting and generalization capabilities.",
author = "D. Alpsan and Ozcan Ozdamar",
year = "1992",
month = "12",
day = "1",
language = "English",
volume = "15",
pages = "83--93",
journal = "Automedica",
issn = "0095-0963",
publisher = "Gordon and Breach Science Publishers",
number = "1",

}

TY - JOUR

T1 - Auditory brainstem evoked potential classification for threshold detection by neural networks. II. Effects of input coding, training set size and composition and network size on performance

AU - Alpsan, D.

AU - Ozdamar, Ozcan

PY - 1992/12/1

Y1 - 1992/12/1

N2 - A feedforward neural network with one hidden layer is applied to the problem of brainstem auditory evoked potential classification. Network performances were tested separately both on subject-dependent samples (drawn from the same subjects from which the training set was derived) and on subject-independent samples (drawn from subjects from which no data was included in the training set), and compared. The results indicate that increasing the training set size improves performance, and human selected training sets give better results than randomly selected sets. Different encoding schemes of signal representation yielded a wide range of correct recognition rates. A spectro-temporal representation was very successful in terms of generalization to novel input and for producing high recognition rates for both classification categories. Although the networks were overly complex and trained in the memorization mode, they displayed feature extracting and generalization capabilities.

AB - A feedforward neural network with one hidden layer is applied to the problem of brainstem auditory evoked potential classification. Network performances were tested separately both on subject-dependent samples (drawn from the same subjects from which the training set was derived) and on subject-independent samples (drawn from subjects from which no data was included in the training set), and compared. The results indicate that increasing the training set size improves performance, and human selected training sets give better results than randomly selected sets. Different encoding schemes of signal representation yielded a wide range of correct recognition rates. A spectro-temporal representation was very successful in terms of generalization to novel input and for producing high recognition rates for both classification categories. Although the networks were overly complex and trained in the memorization mode, they displayed feature extracting and generalization capabilities.

UR - http://www.scopus.com/inward/record.url?scp=0026962635&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0026962635&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0026962635

VL - 15

SP - 83

EP - 93

JO - Automedica

JF - Automedica

SN - 0095-0963

IS - 1

ER -