This paper presents a new compensation mechanism to be used with a Partial Logistic Artificial Neural Network for Competing Risks with Automatic Relevance Determination (PLANN-CR-ARD) and tested comprehensibly on a real breast cancer dataset with excellent convergence properties and numerical stability for the non-linear model. The Model Selection is implemented for the PLANN-CR-ARD model, benefiting from a scaling of the prior error term which together with the data error term forms the total error function that is optimized. The PLANN-CR-ARD proves to be an excellent prognostic tool that can be used in regression analysis tasks such as the survival analysis of cancer datasets.

Model Selection with PLANN-CR-ARD / C.T.C. Arsene, P.J. Lisboa, E. Biganzoli - In: Advances in computational intelligence : 11. International Work-Conference on Artificial Neural Networks : IWANN 2011 : Torremolinos-Málaga, Spain, June 8-10, 2011 : proceedings. 2 / [a cura di] J. Cabestany, I. Rojas, G. Joya. - [s.l] : Springer, 2011. - ISBN 978-3-642-21497-4. - pp. 210-219 (( Intervento presentato al 11. convegno International Work-Conference on Artificial Neural Networks tenutosi a Torremolinos-Málaga nel 2011 [10.1007/978-3-642-21498-1_27].

Model Selection with PLANN-CR-ARD

E. Biganzoli
Ultimo
2011

Abstract

This paper presents a new compensation mechanism to be used with a Partial Logistic Artificial Neural Network for Competing Risks with Automatic Relevance Determination (PLANN-CR-ARD) and tested comprehensibly on a real breast cancer dataset with excellent convergence properties and numerical stability for the non-linear model. The Model Selection is implemented for the PLANN-CR-ARD model, benefiting from a scaling of the prior error term which together with the data error term forms the total error function that is optimized. The PLANN-CR-ARD proves to be an excellent prognostic tool that can be used in regression analysis tasks such as the survival analysis of cancer datasets.
Settore MED/01 - Statistica Medica
Book Part (author)
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/2434/193541
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 9
  • ???jsp.display-item.citation.isi??? 7
social impact