• DocumentCode
    2320007
  • Title

    Evolving a social fabric to fit and epidemic profile

  • Author

    Ashlock, Daniel ; Shiller, Elisabeth

  • Author_Institution
    Dept. of Math. & Stat., Univ. of Guelph, Guelph, ON, Canada
  • fYear
    2012
  • fDate
    9-12 May 2012
  • Firstpage
    363
  • Lastpage
    370
  • Abstract
    Epidemic models often incorporate contact networks along which the disease can be passed. This study follows up on an earlier one which evolved full general contact networks. This study uses an evolvable network representation inspired by the idea of a social fabric. The resulting representation is based on selecting overlapping groups of agents that interact as if they are well mixed. The groups in this representation are intended to represent groups that are, in fact, well mixed such as schools, families, or workplaces. The new representation permits a substantial improvement in the speed with which a contact model can be fit to an epidemic profile. There is a cost in the form of additional model parameters that must be tuned. A parameter setting study is performed for a simple epidemic profile, providing proof of concept for the evolvable social fabric representation. A number of potential improvements and directions for future work are outlined.
  • Keywords
    complex networks; epidemics; evolutionary computation; epidemic models; epidemic profile; evolvable network representation; evolvable social fabric representation; full general contact networks; overlapping agent groups; parameter setting study; social fabric evolution; Computational modeling; Contracts; Diseases; Educational institutions; Evolutionary computation; Fabrics; Social network services; epidemiology; evolutionary computation; evolvable network; network representation; social fabric;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), 2012 IEEE Symposium on
  • Conference_Location
    San Diego, CA
  • Print_ISBN
    978-1-4673-1190-8
  • Type

    conf

  • DOI
    10.1109/CIBCB.2012.6217253
  • Filename
    6217253