[Back]

R. Viertl:
"Fuzzy Information and Statistics";
Keynote Lecture: IJCCI 2012 4th International Joint Conference on Computational Intelligence, Barcelona / Spain (invited); 2012-10-05 - 2012-10-07.

Data in applications are frequently not precise numbers, vectors, and symbols but more or less non-precise also called fuzzy. All measurement results of continuous quantities are subject to fuzziness. This makes it necessary to describe data by fuzzy models before analyzing them. There are different kinds of uncertainty related to data analysis and especially statistics. The most important are variability, imprecision of data, model uncertainty, and uncertainty of a-priori information. Whereas variability is modeled since a long time by probability models, the quantitative mathematical description of imprecision by so-called fuzzy models was done more recently. Especially in statistical data analysis in a Bayesian context also a-priori information is best modeled by so-called fuzzy probability distributions. Examples of non-precise data and related fuzzy models as well as adapted statistical data analysis will be given in the contribution.

www.ijcci.org