Work

Fusion Regression

Year: 2025

Type: article

Abstract: In recent years, various regression methods have been studied in the literature. Although these methods have shown success in different applications, there is no consensus on which one is the best. Different regressors can produce significantly different prediction results when applied to datasets with varying properties. In this paper, we propose Fusion Regression (FuR), a novel approach that combines the predictions of multiple regressors to leverage their complementary views. FuR concatenates the predictions of regressors to create a new feature space and employs a re-ranking scheme for improved accuracy. Our experiments, conducted on 10 datasets with varying properties (such as size and dimension), show that FuR leads to performance gains of up to 20% compared to the best baseline regressor and up to 16% compared to the recently proposed Regression by Re-ranking method.

Source: Pattern Recognition Letters

DOI: 10.1016/j.patrec.2025.03.027

Authors: Filipe Marcel Fernandes Gonçalves, Daniel Carlos Guimarães Pedronette, Ricardo da Silva Torres

Institutions: Universidade Estadual de Campinas (UNICAMP), Universidade Estadual Paulista (Unesp)

Language: English

Topic: Machine Learning and Data Classification

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $2,500

Funders: Fundação de Amparo à Pesquisa do Estado de São Paulo, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Petrobras