On the Optimum Linear Soft Fusion of Classifiers

We present new analytical developments that contribute to a better understanding of the (soft) fusion of classifiers. To this end, we propose an optimal linear combiner based on a minimum mean-square-error class estimation approach. This solution allows us to define a post-fusion mean-square-error i...

Full description

Saved in:
Bibliographic Details
Main Authors: Luis Vergara, Addisson Salazar
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/9/5038
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850278667533418496
author Luis Vergara
Addisson Salazar
author_facet Luis Vergara
Addisson Salazar
author_sort Luis Vergara
collection DOAJ
description We present new analytical developments that contribute to a better understanding of the (soft) fusion of classifiers. To this end, we propose an optimal linear combiner based on a minimum mean-square-error class estimation approach. This solution allows us to define a post-fusion mean-square-error improvement factor relative to the best fused classifier. Key elements for this improvement factor are the number of classifiers, their pairwise correlations, the imbalance between their performances, and the bias. Furthermore, we consider exponential models for the class-conditional probability densities to establish the relationship between the classifier’s error probability and the mean square error of the class estimate. This allows us to predict the reduction in the post-fusion error probability relative to that of the best classifier. These theoretical findings are contrasted in a biosignal application for the detection of arousals during sleep from EEG signals. The results obtained are reasonably consistent with the theoretical conclusions.
format Article
id doaj-art-060c6bf4fe834ce592932fee9f09bb77
institution OA Journals
issn 2076-3417
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-060c6bf4fe834ce592932fee9f09bb772025-08-20T01:49:24ZengMDPI AGApplied Sciences2076-34172025-05-01159503810.3390/app15095038On the Optimum Linear Soft Fusion of ClassifiersLuis Vergara0Addisson Salazar1Instituto de Telecomunicaciones y Aplicaciones Multimedia, Universitat Politècnica de València, C/Camino de Vera s/n, 46022 València, SpainInstituto de Telecomunicaciones y Aplicaciones Multimedia, Universitat Politècnica de València, C/Camino de Vera s/n, 46022 València, SpainWe present new analytical developments that contribute to a better understanding of the (soft) fusion of classifiers. To this end, we propose an optimal linear combiner based on a minimum mean-square-error class estimation approach. This solution allows us to define a post-fusion mean-square-error improvement factor relative to the best fused classifier. Key elements for this improvement factor are the number of classifiers, their pairwise correlations, the imbalance between their performances, and the bias. Furthermore, we consider exponential models for the class-conditional probability densities to establish the relationship between the classifier’s error probability and the mean square error of the class estimate. This allows us to predict the reduction in the post-fusion error probability relative to that of the best classifier. These theoretical findings are contrasted in a biosignal application for the detection of arousals during sleep from EEG signals. The results obtained are reasonably consistent with the theoretical conclusions.https://www.mdpi.com/2076-3417/15/9/5038classifierssoft fusionoptimal combinerexpected improvementfusion of biosignals
spellingShingle Luis Vergara
Addisson Salazar
On the Optimum Linear Soft Fusion of Classifiers
Applied Sciences
classifiers
soft fusion
optimal combiner
expected improvement
fusion of biosignals
title On the Optimum Linear Soft Fusion of Classifiers
title_full On the Optimum Linear Soft Fusion of Classifiers
title_fullStr On the Optimum Linear Soft Fusion of Classifiers
title_full_unstemmed On the Optimum Linear Soft Fusion of Classifiers
title_short On the Optimum Linear Soft Fusion of Classifiers
title_sort on the optimum linear soft fusion of classifiers
topic classifiers
soft fusion
optimal combiner
expected improvement
fusion of biosignals
url https://www.mdpi.com/2076-3417/15/9/5038
work_keys_str_mv AT luisvergara ontheoptimumlinearsoftfusionofclassifiers
AT addissonsalazar ontheoptimumlinearsoftfusionofclassifiers