Hierarchical neural networks for survival analysis. NLM AIDSLINE Important note: Information in this article was accurate in 1996. The state of the art may have changed since the publication date.

Click here to return to AIDSLINE main menu
DonateNow
Print this Article


Hierarchical neural networks for survival analysis.

Medinfo. 1995;8 Pt 1:828-32. Unique Identifier : AIDSLINE MED/96174270
Ohno-Machado L; Walker MG; Musen MA; Section on Medical Informatics, Stanford University School of; Medicine, MSOB X-215 Stanford CA 94305-5479, USA.

Abstract: Neural networks offer the potential of providing more accurate predictions of survival time than do traditional methods. Their use in medical applications has, however, been limited, especially when some data is censored or the frequency of events is low. To reduce the effect of these problems, we have developed a hierarchical architecture of neural networks that predicts survival in a stepwise manner. Predictions are made for the first time interval, then for the second, and so on. The system produces a survival estimate for patients at each interval, given relevant covariates, and is able to handle continuous and discrete variables, as well as censored data. We compared the hierarchical system of neural networks with a nonhierarchical system for a data set of 428 AIDS patients. The hierarchical model predicted survival more accurately than did the nonhierarchical (although both had low sensitivity). The hierarchical model could also learn the same patterns in less than half the time required by the nonhierarchical model. These results suggest that the use of hierarchical systems is advantageous when censored data is present, the number of events is small, and time-dependent variables are necessary.
Keywords: Acquired Immunodeficiency Syndrome/MORTALITY California/EPIDEMIOLOGY Cohort Studies Comparative Study Human Longitudinal Studies Models, Statistical *Neural Networks (Computer) Nonlinear Dynamics Sensitivity and Specificity Support, Non-U.S. Gov't Support, U.S. Gov't, P.H.S. *Survival Analysis JOURNAL ARTICLE

KWDacquiredimmunodeficiencysyndrome/mortalitycalifornia/epidemiologycohortstudiescomparativestudyhumanlongitudinalstudiesmodels,statisticalKWDneuralnetworks(computer)nonlineardynamicssensitivityandspecificitysupport,non-uKWDsKWDgov'tsupport,uKWDsKWDgov't,pKWDhKWDs
960630
M9660760

Copyright © 1996 - National Library of Medicine. Reproduced under license with the National Library of Medicine, Bethesda, MD.

AEGiS is a 501(c)3, not-for-profit, tax-exempt, educational corporation. AEGiS is made possible through unrestricted funding from Boehringer Ingelheim, Bridgestone/Firestone Charitable Trust, Bristol-Myers Squibb Company, Elton John AIDS Foundation, Gill Foundation, the National Library of Medicine, Quest Diagnostics, Roche and Trimeris, and donations from users like you. Always watch for outdated information. This article first appeared in 1996. This material is designed to support, not replace, the relationship that exists between you and your doctor.

AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.

Copyright ©1980, 1996. AEGiS. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS, or the party credited as the provider of the content. .