Reduction of Large Training Set by Guided Progressive Sampling: Application to Neonatal Intensive Care Data

François Portet 1 Feng Gao 1 Jim Hunter 1 René Quiniou 2
2 DREAM - Diagnosing, Recommending Actions and Modelling
Inria Rennes – Bretagne Atlantique , IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
Abstract : Although large training sets are supposed to improve the performance of learning algorithms, there are limits to the volume of data such an algorithm can handle. To overcome this problem, we describe an improvement to a progressive sampling method by guiding the construction of a reduced training set. The application of this method to neonatal intensive care data shows that it is possible to reduce a training set to a third of its original size without decreasing performance.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [3 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01006112
Contributor : François Portet <>
Submitted on : Friday, June 13, 2014 - 5:13:47 PM
Last modification on : Friday, November 16, 2018 - 1:39:15 AM
Document(s) archivé(s) le : Saturday, September 13, 2014 - 11:35:10 AM

File

07idamap.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01006112, version 1

Citation

François Portet, Feng Gao, Jim Hunter, René Quiniou. Reduction of Large Training Set by Guided Progressive Sampling: Application to Neonatal Intensive Care Data. in Intelligent Data International Workshop on Analysis in Medicine and Pharmacology (IDAMAP2007), Jul 2007, Amsterdam, Netherlands. pp.1-2. ⟨hal-01006112⟩

Share

Metrics

Record views

439

Files downloads

94