Evaluation of the logarithmic-sensitivity index as a neural network stopping criterion for rare outcomes

Ennett, Colleen M.; Frize, Monique; Scales, N.

doi:10.1109/itab.2003.1222512

Cited by 12 publications

(10 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To be consistent with the previous research, the logarithmic sensitivity index, defined in (1) was still adopted, although a maximum log-likelihood criterion could be employed. The logarithmic-sensitivity index attempts to achieve optimal sensitivity and specificity of a classifier while slightly favoring higher sensitivity [11]. log -sensitivity -index = -sensitivity n * log10 (1 -sensitivity * specificity) (1)…”

Section: Introductionmentioning

confidence: 99%

Predicting Probability of Mortality in the Neonatal Intensive Care Unit

Zhou¹,

Frize²

2006

2006 International Conference of the IEEE Engineering in Medicine and Biology Society

View full text Add to dashboard Cite

Artificial neural networks can be trained to predict outcomes in a neonatal intensive care unit (NICU). This paper expands on past research and shows that neural networks trained by the maximum likelihood estimation criterion will approximate the 'a posteriori probability' of NICU mortality. A gradient ascent method for the weight update of three-layer feed-forward neural networks was derived. The neural networks were trained on NICU data and the results were evaluated by performance measurement techniques, such as the Receiver Operating Characteristic Curve and the Hosmer-Lemeshow test. The resulting models applied as mortality prognostic screening tools are presented.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting Probability of Mortality in the Neonatal Intensive Care Unit

Zhou¹,

Frize²

2006

2006 International Conference of the IEEE Engineering in Medicine and Biology Society

View full text Add to dashboard Cite

show abstract

“…sensitivity and specificity of the neural network (Ennett, Frize, & Scales, 2003). The logarithmic sensitivity index, defined in Equation 4-4, expands towards infinity when both sensitivity and specificity are 1, but is slightly weighted in favor of sensitivity.…”

Section: 5 L O G a R I T H M I C S E N S I T I V I T Y I N D E Xmentioning

confidence: 99%

“…The logarithmic sensitivity index, defined in Equation 4-4, expands towards infinity when both sensitivity and specificity are 1, but is slightly weighted in favor of sensitivity. The degree of weighting (or bias) towards sensitivity is controlled by the exponent n (Ennett, Frize, & Scales, 2003), which is equal to 0.75 for the experiments related to this thesis.…”

Section: 5 L O G a R I T H M I C S E N S I T I V I T Y I N D E Xmentioning

confidence: 99%

An integrated hybrid data mining systems for preterm birth risk assessment based on a 'semantic web services for healthcare' framework

Catley¹

View full text Add to dashboard Cite

a n In t e g r a t e d H y b r id D a t a M in in g Sy s t e m f o r P r e t e r m B ir t h R is k A ss e s s m e n t B a se d o n a 'Se m a n t ic W e b Ser v ic e s f o r H e a l t h c a r e ' Fr a m e w o r k by Christina Anne Catley, M.A.Sc., B.Eng. A thesis submitted to The Faculty of Graduate Studies and Research in partial fulfillment of the requirements for the degree ofD o c t o r o f Ph il o s o p h y IN ELECTRICAL ENGINEERING

show abstract

“…The logarithmic sensitivity index is typically the chosen stopping criterion for MIRG's ANN research framework. The key attribute with this performance measure is that it maximises both sensitivity and specificity, but is more weighted towards the sensitivity o f the experiment to encourage the ANN to better predict the less frequent outcome (hospitalisation) (Ennett 2002). The functional representation o f the logarithmic sensitivity index is found below (equation 3.7): Index = -Sensitivity" * log(l -Sensitivity * Specificity) (3.7)…”

Section: Figure 12 Sample Receiver Operator Characteristics Curve (Tmentioning

confidence: 99%

“…The ANN learns from one epoch to the next via nine different control parameters that modify the weights and biases o f each node in the network to encourage optimal performance from the network. Table 12 lists the nine parameters and their function in the ANN's learning as explained by Ennett (2002) and Table 13 lists their default ranges found to be ideal by previous researchers in MIRG (Ennett 2002, Rybchynski 2005.…”

Section: Figure 12 Sample Receiver Operator Characteristics Curve (Tmentioning

confidence: 99%

Adverse medical event reduction technologies: identification of primary risk factors of adverse medical events using artificial neural networks

Rideout¹

View full text Add to dashboard Cite

Evaluation of the logarithmic-sensitivity index as a neural network stopping criterion for rare outcomes

Cited by 12 publications

References 7 publications

Predicting Probability of Mortality in the Neonatal Intensive Care Unit

Predicting Probability of Mortality in the Neonatal Intensive Care Unit

An integrated hybrid data mining systems for preterm birth risk assessment based on a 'semantic web services for healthcare' framework

Adverse medical event reduction technologies: identification of primary risk factors of adverse medical events using artificial neural networks

Contact Info

Product

Resources

About