Please use this identifier to cite or link to this item:
https://hdl.handle.net/10216/99872
Author(s): | João Mendes Moreira Alípio Mário Jorge Carlos Soares Jorge Freire de Sousa |
Title: | Ensemble learning: A study on different variants of the dynamic selection approach |
Issue Date: | 2009 |
Abstract: | Integration methods for ensemble learning can use two different approaches: combination or selection. The combination approach (also called fusion) consists on the combination of the predictions obtained by different models in the ensemble to obtain the final ensemble predication. The selection approach selects one (or more) models from the ensemble according to the prediction performance of these models on similar data from the validation set. Usually, the method to select similar data is the k-nearest neighbors with the Euclidean distance. In this paper we discuss other approaches to obtain similar data for the regression problem. We show that using similarity measures according to the target values improves results. We also show that selecting dynamically several models for the prediction task increases prediction accuracy comparing to the selection of just one model. |
Subject: | Inteligência artificial, Ciências da computação e da informação Artificial intelligence, Computer and information sciences |
Scientific areas: | Ciências exactas e naturais::Ciências da computação e da informação Natural sciences::Computer and information sciences |
URI: | https://hdl.handle.net/10216/99872 |
Source: | Machine Learning and Data Mining in Pattern Recognition |
Document Type: | Artigo em Livro de Atas de Conferência Internacional |
Rights: | restrictedAccess |
Appears in Collections: | FCUP - Artigo em Livro de Atas de Conferência Internacional FEP - Artigo em Livro de Atas de Conferência Internacional FEUP - Artigo em Livro de Atas de Conferência Internacional |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
61869.pdf Restricted Access | 269.43 kB | Adobe PDF | Request a copy from the Author(s) |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.