• Title of article

    Maximal confidence intervals of the interval-valued belief structure and applications

  • Author/Authors

    Zhi-gang Su، نويسنده , , Peihong Wang، نويسنده , , Xiang-jun Yu، نويسنده , , Zhen-zhong Lv، نويسنده ,

  • Issue Information
    روزنامه با شماره پیاپی سال 2011
  • Pages
    22
  • From page
    1700
  • To page
    1721
  • Abstract
    An unsolved problem in the interval-valued belief structure is the existence of maximal confidence intervals, referring to the credible and maximal interval-valued beliefs assigned to focal elements, in which every point can be reached by a combination rule. In this study, the existence of maximal confidence intervals of the interval-valued belief structure was investigated and validated. In addition, some applications of the maximal confidence interval in constructing valid and normalized interval-valued belief structures are discussed. A numerical experiment was conducted to illustrate the existence of maximal confidence intervals. Several propositions and a theorem were further proposed to demonstrate the validity of the maximal confidence interval. By using the concept of the maximal confidence interval, a procedure was proposed to construct valid and normalized interval-valued belief structures, and some meaningful guidance for assigning interval-valued beliefs was also suggested to the experts. Finally, a series of well-designed examples was presented as applications of the maximal confidence interval, the proposed procedure and the suggestions. They indicated that valid and normalized interval-valued belief structures can be achieved accurately and efficiently in practical applications with the proposed procedure and guidance on the basis of the concept of the maximal confidence interval.
  • Keywords
    Dempster–Shafer’s theory of evidence , Combination of interval-valued belief , Interval-valued belief structure , Imprecise probability mass , Maximal confidence interval
  • Journal title
    Information Sciences
  • Serial Year
    2011
  • Journal title
    Information Sciences
  • Record number

    1214338