| Texto completo | |
| Autor(es): |
Miranda, Thiago Zafalon
;
Sardinha, Diorge Brognara
;
Basgalupp, Marcio Porto
;
Jin, Yaochu
;
Cerri, Ricardo
Número total de Autores: 5
|
| Tipo de documento: | Artigo Científico |
| Fonte: | PROCEEDINGS OF THE 2020 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE COMPANION, GECCO 2020 COMPANION; v. N/A, p. 2-pg., 2020-01-01. |
| Resumo | |
Multi-label classification consists in classifying an instance into two or more classes simultaneously. Recently, the interest in interpretable classification models have grown, partially as a consequence of regulations such as the General Data Protection Regulation. In this context, we propose a multi-objective evolutionary algorithm that generates multiple rule-based multi-label classification models, allowing users to choose among models that offer different compromises between predictive power and interpretability. The most important contributions of this work are: the generated models are based on sets (unordered collection) of rules and the rule creation mechanism employs a conflict avoidance algorithm which guarantees that all rules within a model are consistent with each other. We conducted experiments on synthetic and real-world datasets and compared our results with state-of-the-art algorithms in terms of predictive performance (F-Score) and interpretability (model size), and demonstrate that our best models had comparable F-Score and smaller model sizes. (AU) | |
| Processo FAPESP: | 16/50457-5 - Multi-objective evolutionary methods for hierarchical and multi-label classification |
| Beneficiário: | Ricardo Cerri |
| Modalidade de apoio: | Auxílio à Pesquisa - Regular |