Improvement of the Interpretability of Fuzzy Rule Based Systems: Quantifiers, Similarities and Aggregators

Abstract : The automatic generation of fuzzy rules is a well-known task based either on the construction of a fuzzy decision tree or on the direct generation of fuzzy rules (e.g. association rules, gradual rules, or fuzzy summaries). In this paper, fuzzy rules obtained from fuzzy decision trees are compared to fuzzy summaries. In both approaches, the knowledge is presented to users using linguistic terms. In this framework, means to improve the interpretability of such fuzzy rules are necessary. The aim of this paper is to reinforce each method by taking advantage of the qualities of the other one. Several methods are proposed, mainly based on the use of fuzzy quantifiers, similarities and aggregations.
Document type :
Book sections
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01533257
Contributor : Lip6 Publications <>
Submitted on : Tuesday, June 6, 2017 - 11:25:14 AM
Last modification on : Thursday, March 21, 2019 - 2:43:15 PM

Links full text

Identifiers

Citation

Anne Laurent, Christophe Marsala, Bernadette Bouchon-Meunier. Improvement of the Interpretability of Fuzzy Rule Based Systems: Quantifiers, Similarities and Aggregators. Modelling with Words, 2873, Springer, pp.102-123, 2003, Lecture Notes in Artificial Intelligence, ⟨10.1007/978-3-540-39906-3_6⟩. ⟨hal-01533257⟩

Share

Metrics

Record views

54