On the study of nearest neighbor algorithms for prevalence estimation in binary problems

PATTERN RECOGNITION

卷:46

Article

Barranquero, Jose¹  Gonzalez, Pablo¹  Diez, Jorge^1,2  Jose del Coz, Juan^1,2 

[1] Univ Oviedo, Ctr Artificial Intelligence, Gijon 33204, Spain

[2] Univ Oviedo, Dept Comp Sci, Gijon 33204, Spain

关键词: Quantification;    Prevalence estimation;    Nearest neighbor;    Methodology;   

DOI  :  10.1016/j.patcog.2012.07.022

来源: Elsevier

【 摘 要 】

This paper presents a new approach for solving binary quantification problems based on nearest neighbor (NN) algorithms. Our main objective is to study the behavior of these methods in the context of prevalence estimation. We seek for NN-based quantifiers able to provide competitive performance while balancing simplicity and effectiveness. We propose two simple weighting strategies, PWK and PWW alpha, which stand out among state-of-the-art quantifiers. These proposed methods are the only ones that offer statistical differences with respect to less robust algorithms, like CC or AC. The second contribution of the paper is to introduce a new experiment methodology for quantification. (C) 2012 Elsevier Ltd. All rights reserved.

【 授权许可】

Free