Enhancing evolutionary fuzzy systems for multi-class problems: Distance-based relative competence weighting with truncated confidences (DRCW-TC)
详细信息 查看全文
- 作者：Alberto Ferná ; ndeza ; alberto.fernandez@ujaen.es" class="auth_mail" title="E-mail the corresponding author ; Mikel Elkanob ; mikel.elkano@unavarra.es" class="auth_mail" title="E-mail the corresponding author ; Mikel Galarb ; mikel.galar@unavarra.es" class="auth_mail" title="E-mail the corresponding author ; José ; Antonio Sanzb ; joseantonio.sanz@unavarra.es" class="auth_mail" title="E-mail the corresponding author ; Saleh Alshomranid ; f ; sshomrani@kau.edu.sa" class="auth_mail" title="E-mail the corresponding author ; Humberto Bustinceb ; c ; bustince@unavarra.es" class="auth_mail" title="E-mail the corresponding author ; Francisco Herrerae ; f ; herrera@decsai.ugr.es" class="auth_mail" title="E-mail the corresponding author
- 关键词：Multi-class classification ; Pairwise learning ; One-vs-One ; Classifier selection ; Evolutionary fuzzy systems ; Fuzzy rule based classification systems
- 刊名：International Journal of Approximate Reasoning
- 出版年：2016
- 出版时间：June 2016
- 年：2016
- 卷：73
- 期：Complete
- 页码：108-122
- 全文大小：792 K
文摘
- class="listitem" id="list_ls0010">
- class="label">•
The problem of non-competence for pairwise learning is addressed.
- class="label">•
A new methodology, based on truncation of the confidence degrees, is proposed.
- class="label">•
The properties of Fuzzy Rule Based Classification Systems are taken into account in the design of this novel model.
- class="label">•
A distance-based tuning is carried out to adapt the score-matrix of the One-vs-One procedure.
- class="label">•
Experimental results versus the state-of-the-art show the goodness of this approach.