Title: Fuzzy Statistical Analysis
Description:
A basic problem, at the present stage of the
Information society, is how to manage the cognitive process while taking into
account its intrinsic features of uncertainty, including imprecision and
vagueness. This has both theoretical and practical implications in Technology,
Economics, Bio-medicine, etc. Traditional Statistics has developed tools and
procedures for coping with this problem, assuming that uncertainty is basically
due to random mechanisms appropriately handled by means of Probability. The
theory of Fuzzy Sets and its generalization to what may be called "Fuzzy
thinking" has widened the scope of Statistics enabling us to deal with more
general sources of uncertainty such as vagueness and imprecision as referred to
both empirical data and/or models for data analysis.
On the other hand, the need for statistical methods capable of managing complex
problems exploiting heterogeneous information (diversity of definitions and
measurements, different degrees of imprecision and confidence with reference to
both data and assumptions) has been greatly increased by the availability of
large data bases and information systems in almost every field of activity. In
recent years, many of the ideas and techniques developed by the "Fuzzy Community"
have been translated into the statistical language, yielding new methodological
proposals, especially (but not exclusively) in the areas of Regression and
Cluster Analysis. At the same time, several suggestions have put forward,
concerning the use of fuzzy random variables and of the traditional
probabilistic machinery for extending the scope of classical (crisp) inferential
methods (e.g. the general linear model) to cover fuzzy data. In spite of the
impact of this growing literature, there is relevant room for further
developments in several directions, including foundations, methodology and
applications of the Fuzzy approach to Statistics.
Focus:
This track is meant to stimulate and present
contributions and all of above directions. Foundational issues may include:
- Use of Possibility Theory in Statistics
- Construction and utilization of Fuzzy Probabilistic models in statistical
inference
- Relationship between Conditional Probability and Fuzzy Information
- Comparison between Fuzzy and Traditional (crisp) Statistical Methods
- Fuzzy approach vs Interval Analysis approach Methodological
Contributors may encompass such field as:
- Regression (possibilistic, least squares, or mixed approaches)
- Classification (cluster and discriminant analyses)
- Latent structure analysis and scaling techniques
- Inferential procedures for the analysis of dependence and association.
Applications of Fuzzy Statistical methods may cover various domains such as:
- Biomedicine
- Finance and Economics
- Social sciences and Psychology
- Environmental sciences
- Technology
The track will offer a forum for discussing and comparing different viewpoints
and proposals on a critical junction of the new Information Science, located at
the crossroads of so far separated pathways of knowledge and research: cognitive
sciences, artificial intelligence, engineering, informatics, mathematics,
philosophy of science, statistics, and others.
Co-Chairs:
Renato Coppi
Department of Statistics
University of Rome
P.le Aldo Moro, 5
00185 Rome
Italy
E-mail: renato.coppi@uniroma1.it
|
Maria Angeles Gil
Department of Statistics
University of Oviedo
C/Calvo Sotelo,s/n
33007 Oviedo
Spain
E-mail: angeles@pinon.ccu.uniovi.es |
|
|
|