Constructing Predictive Belief Functions from Continuous Sample Data Using Confidence Bands
Astride Aregui · 2007
We consider the problem of quantifying our belief in future values of a random variable X with unknown distribution PX, based on the observation of a random sample from the same distribution. The adopted uncertainty representation framework is the Transferable Belief Model, a subjectivist interpretation of belief function theory. In a previous paper, the concept of predictive belief function at a given confidence level was introduced, and it was shown how to build such a function when X is discrete. This work is extended here to the case where X is a continuous random variable, based on step or continuous confidence bands.