Optimizing a neural network to improve classification performance

D. Alpson, M. Towsey, D. N. Ghista, O. Ozdamar, A. Tsoi

Research output: Contribution to journalArticlepeer-review

12 Scopus citations


The paper describes the use of feedforward neural networks in conjunction with back propagation for automating the classification of biomedical signals, particularly in determining hearing threshold from Brainstem Auditory Evoked Potentials (BAEPs). The neural networks were fine tuned for gain, momentum, batch size, and hidden layer size in order to maximize generalization resulting in the reduction of false negative classifications, with only a small sacrifice in learning speed.

Original languageEnglish (US)
Pages (from-to)465-471
Number of pages7
JournalIEEE Engineering in Medicine and Biology Magazine
Issue number4
StatePublished - 1994

ASJC Scopus subject areas

  • Biomedical Engineering


Dive into the research topics of 'Optimizing a neural network to improve classification performance'. Together they form a unique fingerprint.

Cite this