Facebook Instagram Twitter RSS Feed PodBean Back to top on side

Monte-Carlo estimations for belief functions

In: Tatra Mountains Mathematical Publications, vol. 16, no. 2
Ivan Kramosil
Detaily:
Rok, strany: 1999, 339 - 358
O článku:
Belief functions serve as the main numerical characteristics of the degrees of uncertainty in the so called Dempster-Shafer theory. Even if belief functions do not possess, in general, the properties of probability measures, particular values of belief functions can be defined by probabilities with which an appropriately defined generalized set-valued random variable takes values satisfying certain quite natural properties. Consequently, the values of belief functions can be estimated, using the most elementary principles of Monte-Carlo methods, by relative frequencies of successes in series of statistically independent and identically distributed random samples. An algorithm based on this idea is proposed and its asymptotical properties as well as its qualitative characteristics in non-asymptotic cases are stated and proved.
Ako citovať:
ISO 690:
Kramosil, I. 1999. Monte-Carlo estimations for belief functions. In Tatra Mountains Mathematical Publications, vol. 16, no.2, pp. 339-358. 1210-3195.

APA:
Kramosil, I. (1999). Monte-Carlo estimations for belief functions. Tatra Mountains Mathematical Publications, 16(2), 339-358. 1210-3195.