Learning of FCMs with causal links represented via fuzzy triangular numbers

M. Furkan Dodurka, Atakan Sahin, Engin Yesil, Leon Urbas

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

8 Citations (Scopus)

Abstract

In this paper, learning of the FCMs represented using triangular fuzzy numbers (TFNs) in their weight matrices is studied. For this aim a population based novel learning approach is proposed. In the proposed algorithm, BB-BC optimization method is preferred because of its fast convergence capability. Moreover, this proposed approach involves concept by concept (CbC) learning to increase the accuracy of the learning of FCMs. Two different tests are realized as case studies for investigating the performance of the learning approach. For the first test, the learning capability of the algorithm is examined and for the second test the performance of generalization capability is investigated. The tests, which are presented via tables and figures, show that learning approach is successful for learning of FCMs with TFNs. Furthermore, from the case study it can be seen that the uncertain information can be represented and interpreted by the proposed FCM design methodology in a more efficient way.

Original languageEnglish
Title of host publication2015 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE 2015)
Place of PublicationPiscataway, NJ.
PublisherIEEE
Number of pages8
ISBN (Print)9781467374286
DOIs
Publication statusPublished - 25 Nov 2015
EventIEEE International Conference on Fuzzy Systems, FUZZ-IEEE 2015 - Istanbul, Turkey
Duration: 2 Aug 20155 Aug 2015

Conference

ConferenceIEEE International Conference on Fuzzy Systems, FUZZ-IEEE 2015
Country/TerritoryTurkey
CityIstanbul
Period2/08/155/08/15

Keywords

  • causal Links
  • fuzzy cognitive maps
  • learning
  • reasoning
  • triangular fuzzy numbers
  • weight matrix

Fingerprint

Dive into the research topics of 'Learning of FCMs with causal links represented via fuzzy triangular numbers'. Together they form a unique fingerprint.

Cite this