Document Server@UHasselt >
Research >
Research publications >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1942/22997

Title: Partitive Granular Cognitive Maps to Graded Multilabel Classification
Authors: Nápoles Ruiz, Gonzalo
Falcon, Rafael
Papageorgiou, Elpiniki
Bello, Rafael
Vanhoof, Koen
Issue Date: 2016
Publisher: IEEE
Citation: 2016 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), IEEE,p. 64-71
Series/Report: IEEE World Congress on Computational Intelligence
Abstract: In a multilabel classification problem, each object gets associated with multiple target labels. Graded multilabel classification (GMLC) problems go a step further in that they provide a degree of association between an object and each possible label. The goal of a GMLC model is to learn this mapping while minimizing a certain loss function. In this paper, we tackle GMLC problems from a Granular Computing perspective for the first time. The proposed schemes, termed as partitive granular cognitive maps (PGCMs), lean on Fuzzy Cognitive Maps (FCMs) whose input concepts represent cluster prototypes elicited via Fuzzy C-Means whereas the output concepts denote the set of existing labels. We consider three different linkages between the FCM’s input and output concepts and learn the causal connections (weight matrix) through a Particle Swarm Optimizer (PSO). During the exploitation phase, the membership grades of a test object to each fuzzy cluster prototype in the PGCM are taken as the initial activation values of the recurrent network. Empirical results on 16 synthetically generated datasets show that the PGCM architecture is capable of accurately solving GMLC instances.
Notes: Napoles, G (reprint author), Hasselt Univ Campus Diepenbeek, BE-3590 Diepenbeek, Belgium. gonzalo.napoles@student.uhasselt.be; rfalcon@uottawa.ca; epapageorgiou@mail.teiste.gr; rbellop@uclv.edu.cu; koen.vanhoof@uhasselt.be
URI: http://hdl.handle.net/1942/22997
DOI: 10.1109/FUZZ-IEEE.2016.7737848
ISI #: 000392150700189
ISBN: 9781509006250
ISSN: 1544-5615
Category: C1
Type: Proceedings Paper
Appears in Collections: Research publications

Files in This Item:

Description SizeFormat
Published version1.31 MBAdobe PDF
Peer-reviewed author version1.13 MBAdobe PDF

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.