|
Activity Number:
|
342
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305368 |
|
Title:
|
Gene Expression (Microarray) Analysis by Neural Networks
|
|
Author(s):
|
David Booth*+ and David Zhu and Richard Geoke and David Baker and James Hamburg
|
|
Companies:
|
Kent State University and Kent State University and Kent State University and Kent State University and Kent State University
|
|
Address:
|
Management and IS Department, Kent, OH, 44242,
|
|
Keywords:
|
microarray ; gene expression ; leukemia ; neural networks ; cross validation ; cancer
|
|
Abstract:
|
Three sets of optimal leukemia class predictors (genes) were obtained by three methods from three authors from the same set of data. We tested these optimal sets using back propagation neural networks (which were not used by the original authors) with three-fold cross-validation and leave-one-out cross-validation. We found that the predictor sets performed poorer with the neural networks than with the original methods, though not in all cases. We discuss this result and suggest methods for possibly taking advantage of this finding.
|
