Negative examples, which are required for most machine learning methods to infer new predictions, are rarely directly recorded in several real world databases for classification problems. A variety of heuristics for the choice of negative examples have been proposed, ranging from simply under-sampling non positive instances, to the analysis of class taxonomy structures. Here we propose an efficient strategy for selecting negative examples designed for Hopfield networks which exploits the clustering properties of positive instances. The method has been validated on the prediction of protein functions of a model organism.
Selection of negative examples for node label prediction through fuzzy clustering techniques / M. Frasca, D. Malchiodi (SMART INNOVATION, SYSTEMS AND TECHNOLOGIES). - In: Advances in Neural Networks : Computational Intelligence for ICT / [a cura di] S. Bassis, A. Esposito, F. C. Morabito, E. Pasero. - Prima edizione. - [s.l] : Springer International, 2016. - ISBN 9783319337463. - pp. 67-76 (( Intervento presentato al 25. convegno WIRN tenutosi a Vietri sul mare nel 2015 [10.1007/978-3-319-33747-0_7].
Selection of negative examples for node label prediction through fuzzy clustering techniques
M. Frasca;D. Malchiodi
2016
Abstract
Negative examples, which are required for most machine learning methods to infer new predictions, are rarely directly recorded in several real world databases for classification problems. A variety of heuristics for the choice of negative examples have been proposed, ranging from simply under-sampling non positive instances, to the analysis of class taxonomy structures. Here we propose an efficient strategy for selecting negative examples designed for Hopfield networks which exploits the clustering properties of positive instances. The method has been validated on the prediction of protein functions of a model organism.File | Dimensione | Formato | |
---|---|---|---|
WIRN15_5.pdf
accesso riservato
Tipologia:
Post-print, accepted manuscript ecc. (versione accettata dall'editore)
Dimensione
296.55 kB
Formato
Adobe PDF
|
296.55 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.