Iranian Journal of Fuzzy Systems
Year:2004 | Volume:1 | Issue:1|Papers Count:7

The paper analyses issues leading to errors in graphic object classifiers. The distance measures suggested in literature and used as a basis in traditional, fuzzy, and Neuro-Fuzzy classifiers are found to be not suitable for classification of non-stylized or fuzzy objects in which the features of classes are much more difficult to recognize because of significant uncertainties in their location and gray-levels. The authors suggest a Neuro-Fuzzy graphic object classifier with modified distance measure that gives better performance indices than systems based on traditional ordinary and cumulative distance measures. The simulation has shown that the quality of recognition significantly improves when using the suggested method.

While sophisticated analytical methods like Morgenstern-Price or finite element methods are available for more realistic analysis of stability of slopes, assessment of the exact values of soil parameters is practically impossible. Uncertainty in the soil parameters arises from two different sources: scatter in data and systematic error inherent in the estimate of soil properties. Hence, stability of a slope should be expressed using a factor of safety accompanied by a reliability index. In this paper, theory of fuzzy sets is used to deal with the uncertain nature of soil parameters and the inaccuracy involved in the analysis simultaneously. Soil parameters are defined using suitable fuzzy sets and the uncertainty inherent in the value of factor of safety is assessed accordingly. It is believed that this approach accounts for the uncertainty in soil parameters more realistically compared to the conventional probabilistic approaches reported in the literature. A computer program is developed that carries out the large amount of calculations required for evaluating the fuzzy factor of safety based on the concept of domain interval analysis. An aggregated fuzzy reliability index (AFRI) is defined and assigned to the calculated factor of safety. The proposed method is applied to a case study and the results are discussed in details. Results from sensitivity analysis describe where the exploration effort or quality control should be concentrated. The advantage of the proposed method lies in its fast calculation speed as well as its ease of data acquisition from experts* opinion through fuzzy sets.

The pioneer work of image compression/reconstruction based on fuzzy relational equations (ICF) and the related works are introduced. The ICF regards an original image as a fuzzy relation by embedding the brightness level into [0,1]. The compression/reconstruction of ICF correspond to the composition/solving inverse problem formulated on fuzzy relational equations. Optimizations of ICF can be consequently deduced based on fuzzy relational calculus, i.e., computation time reduction/improvement of reconstructed image quality are correspond to a fast solving method/finding an approximate solution of fuzzy relational equations, respectively. Through the experiments using test images extracted from Standard Image DataBAse (SIDBA), the effectiveness of the ICF and its optimizations are shown

In applications there occur different forms of uncertainty. The two most important types are randomness (stochastic variability) and imprecision (fuzziness). In modelling, the dominating concept to describe uncertainty is using stochastic models which are based on probability. However, fuzziness is not stochastic in nature and therefore it is not considered in probabilistic models. Since many years the description and analysis of fuzziness is subject of intensive research. These research activities do not only deal with the fuzziness of observed data, but also with imprecision of informations. Especially methods of standard statistical analysis were generalized to the situation of fuzzy observations. The present paper contains an overview about the presentation of fuzzy information and the generalization of some basic classical statistical concepts to the situation of fuzzy data.

The notion of strong arcs in a fuzzy graph was introduced by Bhutani and Rosenfeld in [1] and fuzzy end nodes in the subsequent paper [2] using the concept of strong arcs. In Mordeson and Yao [7], the notion of "degrees" for concepts fuzzified from graph theory were defined and studied. In this note, we discuss degrees for fuzzy end nodes and study further some properties of fuzzy end nodes and fuzzy cut nodes.

The intuitionistic fuzzification of (strong, weak, s-weak) hyper BCK-ideals is introduced, and related properties are investigated. Characterizations of an intuitionistic fuzzy hyper BCK-ideal are established. Using a collection of hyper BCK-ideals with some conditions, an intuitionistic fuzzy hyper BCK-ideal is built.

In this paper, countable compactness and the Lindelof property are defined for L-fuzzy sets, where L is a complete de Morgan algebra. They don*t rely on the structure of the basis lattice L and no distributivity is required in $L$. A fuzzy compact L-set is countably compact and has the Lindelof property. An L-set having the Lindelof property is countably compact if and only if it is fuzzy compact. Many characterizations of countable compactness and the Lindelof property are presented by means of open L-sets and closed L-sets when L is a completely distributive de Morgan algebra.