Wird geladen...

Ensemble learning by data resampling

We investigate ensemble learning methods that construct a classifier ensemble by repeatedly sampling the original training data and building a member classifier from each subsample. We find that the performance of standard Bagging can frequently be improved upon by simple variations of the sampling...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Goebel, Michael
Format: Online
Sprache:en
Veröffentlicht: ResearchSpace@Auckland 2004
Online Zugang:http://hdl.handle.net/2292/996
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!