Self-Validation Methods in Statistical Computing

点击翻译按钮获取中文摘要

This research is concerned with the development of self-validating algorithms for computing probabilities in selected continuous univariate and multivariate probability distributions. Self-validating numerical methods are needed because of the difficulty in assessing the accuracy of results obtained in large-scale computing from standard methods. Toward this end, interval analysis will be used in the development of these algorithms as well as rounded interval arithmetic in an IBM compatible microcomputer.