Research article summary:

Weight-elimination neural networks applied to coronary surgery mortality prediction.

Abstract Extract:
The objective was to assess the effectiveness of the weight-elimination cost function in improving classification performance of artificial neural networks (ANNs) and to observe how changing the a priori distribution of the training set affects network ... (Full abstract text below)

Published 2003Jun in Journal: IEEE Trans Inf Technol Biomed (Language : eng)

Full Pubmed Extract

This information was retrieved, real-time, on your behalf from the public area of the Pubmed website:

1. IEEE Trans Inf Technol Biomed. 2003 Jun;7(2):86-92

Weight-elimination neural networks applied to coronary surgery mortality prediction.

Ennett CM, Frize M

Systems and Computer Engineering Department, Carleton University, Ottawa, ON K1S 5B6, Canada. ennett@canada.com

The objective was to assess the effectiveness of the weight-elimination cost function in improving classification performance of artificial neural networks (ANNs) and to observe how changing the a priori distribution of the training set affects network performance. Backpropagation feedforward ANNs with and without weight-elimination estimated mortality for coronary artery surgery patients. The ANNs were trained and tested on cases with 32 input variables describing the patient's medical history; the output variable was in-hospital mortality (mortality rates: training 3.7%, test 3.8%). Artificial training sets with mortality rates of 20%, 50%, and 80% were created to observe the impact of training with a higher-than-normal prevalence. When the results were averaged, weight-elimination networks achieved higher sensitivity rates than those without weight-elimination. Networks trained on higher-than-normal prevalence achieved higher sensitivity rates at the cost of lower specificity and correct classification. The weight-elimination cost function can improve the classification performance when the network is trained with a higher-than-normal prevalence. A network trained with a moderately high artificial mortality rate (artificial mortality rate of 20%) can improve the sensitivity of the model without significantly affecting other aspects of the model's performance. The ANN mortality model achieved comparable performance as additive and statistical models for coronary surgery mortality estimation in the literature.

PMID : 12834163 [PubMed - Indexed for MEDLINE]

This information is obtained from the National Library of Medicine (NLM). Abstract text and other information may be subject to copyright. Type "NLM copyright" into Google for more information.

Full Author Information

First Name	LastName	Initials
Colleen M	Ennett	CM
Monique	Frize	M

Affiliation: Systems and Computer Engineering Department, Carleton University, Ottawa, ON K1S 5B6, Canada. ennett@canada.com

3rd Party provider links

Click the links below to go to related 3rd party information:

MESH categories and related page links

This article was linked to the MESH categories shown on the left below. The links on the right are related Memletics pages.

Links for this article

For links to places where you can get the full text of this article see links. Note there may be a subscription or fee required for access to the full text.

New! Using similar technology to this site, we have launched find-health-articles.com, targeting over 1 million health research article abstracts.

Here are some articles related to this one (by title keywords):

Evolving neural networks through augmenting topologies.

Effects of a videotape information intervention for spouses on spouse distress and patient recovery from surgery.

Community structure in social and biological networks.

Classification image weights and internal noise level estimation.

Basic mechanisms for graded persistent activity: discrete attractors, continuous attractors, and dynamic representations.

The association of incident dementia with mortality in PD.

Additive effects of cognitive function and depressive symptoms on mortality in elderly community-living adults.

Racial and ethnic disparity in low birth weight in Syracuse, New York.

Prediction of dihydrofolate reductase inhibition and selectivity using computational neural networks and linear discriminant analysis.

Effects of a limited class of nonlinearities on estimates of relative weights.

>>> Get more research article abstracts like these, or browse abstracts.

Keywords in this article:

achieved, additive, affecting, affects, anns, artery, artificial, aspects, assess, averaged, backpropagation, cases, changing, classification, comparable, coronary, correct, cost, created, describing, distribution, effectiveness, elimination, estimated, estimation, feedforward, function, higher, history, hospital, impact, improve, improving, input, lower, medical, model, models, moderately, mortality, networks, neural, normal, objective, observe, other, output, patients, performance, prevalence, priori, rates, results, sensitivity, set, sets, specificity, statistical, surgery, tested, trained, training, variables, weight

Also, see our new free speed reading online course (beta version)