Showing papers on "Uncertain data published in 1986"

Fuzzy subsets applications in o.r. and management

Abstract: In uncertain environment and forecasting the use of fuzzy numbers is a realistic way to know the effects of non measurables hypothesis. An extension to Zero Base Budgeting Method is shown using triangular fuzzy numbers ; it is also a progress for the realism of this very well known and now classical method to select investments. Fuzzy exponential smoothing concerns short range forecasting and analysis of sequence of numerical data. Practically, for novel equipments, and it is quite often the case, it is very difficult and sometimes impossible to know the survival functionas probability law ; a parallel theory using fuzzy subsets and possibility is then very useful. For long range forecasting, the method Delphi from RAND Corp. is one of the most efficient, we present a variant using triangular fuzzy number and analysis by maximal sub-relations of similarity.

A maximum variance model for robust detection and estimation with dependent data

Abstract: A model that accounts for uncertain data dependency is developed by generating a large class of stationary stochastic processes, each with the same univariate distribution. This class can be considered to be a contamination class about the nominal independent and identically distributed (i.i.d.) process distribution. The class is developed specifically for application to robust detector and estimator design based on asymptotic variance. Application of this dependency class leads to an intuitively pleasing result: the minimax variance estimators and the maximin efficacy detectors are the same as obtained using i.i.d, asymptotic estimation and detection theory. Thus our technique generalizes previously obtained robust design results for i.i.d, data to this dependent data case.

Implementing Probabilistic Reasoning

Abstract: General problems in analyzing information in a probabilistic database are considered. The practical difficulties (and occasional advantages) of storing uncertain data, of using it in conventional forward- or backward-chaining inference engines, and of working with a probabilistic version of resolution are discussed. The background for this paper is the incorporation of uncertain reasoning facilities in MRS, a general-purpose expert system building tool.