No Free Lunch for Early Stopping

We show that with a uniform prior on models having the same training error, early stopping at some fixed training error above the training error minimum results in an increase in the expected generalization error.

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	Neural computation 1999-05, Vol.11 (4), p.995-1009
Hauptverfasser:	Cataltepe, Zehra, Abu-Mostafa, Yaser S., Magdon-Ismail, Malik
Format:	Artikel
Sprache:	eng
Schlagworte:	Algorithms Applied sciences Artificial intelligence Computer science control theory systems Exact sciences and technology Learning and adaptive systems Linear Models Nonlinear Dynamics Probability Reproducibility of Results Time Factors
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	We show that with a uniform prior on models having the same training error, early stopping at some fixed training error above the training error minimum results in an increase in the expected generalization error.
ISSN:	0899-7667 1530-888X
DOI:	10.1162/089976699300016557