S -Convex Extrema, Taylor-Type Expansions and Stochastic Approximations (Q4780929)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1833674
Language Label Description Also known as
default for all languages
No label defined
    English
    S -Convex Extrema, Taylor-Type Expansions and Stochastic Approximations
    scientific article; zbMATH DE number 1833674

      Statements

      S -Convex Extrema, Taylor-Type Expansions and Stochastic Approximations (English)
      0 references
      21 November 2002
      0 references
      stochastic ordering
      0 references
      \(s\)-convex ordering
      0 references
      stochastic expansions
      0 references
      stochastic approximations
      0 references
      stop-loss distance
      0 references
      0 references
      This paper deals with the stochastic approximation of risk based either on \(s\)-convex stochastic extrema, or on the canonical moments. Using the probabilistic generalization of Taylor's theorem the author studies \(s\)-convex stochastic orders. Let \(s\) be a positive integer and let \(X\) and \(Y\) be two risks with finite \((s-1)\)th moments. The risk \(X\) is said to precede the risk \(Y\) in the \(s\)-convex order if \(E\varphi(X)\leq E\varphi(Y)\) for all \(\varphi\in G_{s}:= \{x\to\pm x^{k},\;k=1,2,\ldots,s-1;\;x\to(x-t)^{s-1}_{+}, t\in R^{+}\}\). The approximations of \(X\) from the corresponding moment space by either \(s\)-convex minimum or \(s\)-convex maximum are called extremal approximations. In the canonical approximation the author suggests the use of mixture of \(s\)-convex extrema with weights defined in terms of the canonical moments. The upper bounds on the stop-loss distances between a risk \(X\) and its stochastic approximations are obtained. Several actuarial applications of the obtained results are presented. In particular, the accuracy of the bounds furnished by the \(s\)-convex extrema on quantities of whole life premiums and exponential premiums are demonstrated. In addition to the lower and upper bounds on these quantities, the canonical approximations provide a reasonable value for the unknown risk quantity. The author approximates the effective interest rate of whole life insurance from price list, Lundberg's coefficients as well as zero-utility premiums. Also it is shown that the stochastic approximations can be applied to get estimation of mixing distribution in Poisson and negative exponential mixture models.
      0 references

      Identifiers