J. V. de Oliveira, Semantic constraints for membership function optimization, IEEE Trans. FS 19 (1999) 128--138.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Fuzzy logic improves rule based classifiers by allowing the use of overlapping class definitions and improves the interpretability of the results by providing more insight into the decision making process. Fuzzy logic, however, is not a guarantee for interpretability, as was also recognized in [25, 23]. Hence, real e#ort must be made to keep the resulting rule base transparent. The automatic determination of compact fuzzy classifiers rules from data has been approached by several di#erent techniques: neuro fuzzy methods [15] genetic algorithm (GA) based rule selection [11] and fuzzy ....

Valente de Oliveira J. (1999) Semantic constraints for membership function optimization. IEEE Trans. FS 19, 128--138. 18

....etc. Among the wide range of CI techniques, fuzzy logic improves classification and decision support systems by allowing the use of overlapping class definitions and improves the interpretability of the results by providing more insight into the classifier structure and decision making process [9]. In Section 3. a detailed discussion about the use of fuzzy techniques for knowledge representation in classifier systems will be given. Model evaluation criteria: qualitative statements or fit functions of how well a particular pattern (a model and its parameters) meet the goals of the KDD ....

....Traditionally, algorithms to obtain classifiers have focused either on accuracy or interpretability. Recently some approaches to combining these properties have been reported; fuzzy clustering is proposed to derive transparent models in [10] linguistic constraints are applied to fuzzy modeling in [9] and rule extraction from neural networks is described in [11] Hence, to obtain compact and interpretable fuzzy models model reduction algorithms have to be used that will be overviewed in Section 4. Search method: consists of two components: parameter search and model search. Once the model ....

[Article contains additional citation context not shown here]

J. V. de Oliveira, Semantic constraints for membership function optimization, 1EEE Trans. FS 19 (1999) 128 138.

....as possible. Currently, there exist no well established definitions for interpretability of fuzzy systems, mainly due to the subjective nature of such a concept. However, some works have attempted to define objective criteria that facilitate the automatic modeling of interpretable fuzzy systems [56, 175]. The fuzzy system of Figure 1.11 processes information in three stages: the input interface (fuzzifier) the processing stage (inference engine) and the output interface (defuzzifier) The interface deals with linguistic variables and their corresponding labels. These linguistic variables ....

....variables, each with three labels, divide the input space into a grid of nine regions. Normal, orthogonal membership functions. The membership functions of two successive labels must be complementary (i.e. their sum must be equal to one) in their overlapping region, whatever form they have [37,175]. Moreover, in such regions each label must ascend from zero to unity membership values [125, 129] The variables presented in Figures 2.1 and 2.3, satisfy these requirements. Don t care conditions. A fully defined rule base, as that shown in Figure 2.4a, becomes impractical for high dimension ....

J. Valente de Oliveira. Semantic constraint for membership function optimization. IEEE Transactions on Systems, Man, and Cybernetics. Part A: Systems and Humans, 29(1):128--138, January 1999.

....which are a subclass of decision systems. Each example given to a classifier is associated with one out of a limited number of predefined classes. Linguistic interpretability is an important aspect of a knowledge based classifier. Fuzzy logic helps improving the classifier through its semantics [9] that provide insight in the classifier structure and decision making process. Fuzzy logic, however, is not a guarantee for interpretability, as recognized by [9, 5] Real effort must be made to keep the resulting rule base transparent [7, 2, 12] Two main approaches are found in literature: i) ....

....is an important aspect of a knowledge based classifier. Fuzzy logic helps improving the classifier through its semantics [9] that provide insight in the classifier structure and decision making process. Fuzzy logic, however, is not a guarantee for interpretability, as recognized by [9, 5]. Real effort must be made to keep the resulting rule base transparent [7, 2, 12] Two main approaches are found in literature: i) Select a low number of input variables in order to make a compact classifier [7, ii)Make a large set of possible rules and make a useful selection out of these ....

[Article contains additional citation context not shown here]

Valente de Oliveira J. (1999) Semantic constraints for membership function optimization. IEEE Trans. FS 19, 128--138.

....interpretability is also an important aspect which must be taken into account. The first two aspects are often approached by an exhaustive search or educated guesses, while the interpretability aspect is often neglected. Only recently people recognized the importance of all these aspects [2,3], which makes the automatic data based identification of classification systems that are compact, interpretable and accurate, a challenging topic. We propose fuzzy logic rule based classifiers to handle the interpretability aspect. Fuzzy logic helps to improve the interpretability of ....

....www.fmt.vein.hu softcomp (Janos Abonyi ) http: LCEwww.et.tudelft.nl (Johannes A. Roubos) classifiers through its semantics that provide insight in the classifier structure and decision making process. Fuzzy logic, however, is not a guarantee for interpretability, as was also recognized in [2,3]. Real e#ort must be made to keep the resulting rule base transparent [4 6] For this purpose, two main approaches are followed in the literature: i) Selection of a low number of input variables in order to create a compact classifier [4,7] and (ii) Construction of a large set of possible rules ....

[Article contains additional citation context not shown here]

J. V. de Oliveira, Semantic constraints for membership function optimization, IEEE Trans. FS 19 (1999) 128--138.

....as possible. Currently, there exist no well established definitions for interpretability of fuzzy systems, mainly due to the subjective nature of such a concept. However, some works have attempted to define objective criteria that facilitate the automatic modeling of interpretable fuzzy systems [5, 20]. The fuzzy system of Figure 2 processes information in three stages: the input interface (fuzzifier) the processing stage (inference engine) and the 8 Carlos Andres Pena Reyes and Moshe Sipper [mg dL] 1 P =400 2 P =1000 3 P =200 250 0.75 0.25 0 1 Membership Normal Very High High ....

....does not guarantee the semantic integrity of each variable. More conditions are necessary. Normal, orthogonal membership functions. The membership functions of two successive labels must be complementary (i.e. their sum must be equal to one) in their overlapping region, whatever form they have [4,20]. Moreover, in such regions each label must ascend from zero to unity membership values [13, 14] The variables presented in Figures 3 and 5, satisfy these requirements. Don t care conditions. A fully defined rule base, as that shown in Figure 6a, becomes impractical for high dimension systems. ....

J. Valente de Oliveira. Semantic constraint for membership function optimization. IEEE Transactions on Systems, Man, and Cybernetics. Part A: Systems and Humans, 29(1):128--138, January 1999.

....like fuzzy clustering [5] 6] neural June 5, 2000 DRAFT 3 networks [7] 8] statistical information criteria [9] Kalman filters [9] hill climbing [8] and even fuzzy expert control of the GA s operators [10] to mention some. This has resulted in many complex algorithms and, as recognized in [11] and [12] often the transparency and compactness of the resulting rule base is not considered to be of importance. In such cases, the fuzzy model becomes a black box, and one can question the rationale for applying fuzzy modeling instead of other techniques like, e.g. neural networks. In the ....

....the membership functions derived from data have simple shapes and are well separated, they can still be assigned meaningful linguistic labels by the domain experts. he initial rule base constructed by fuzzy clustering typically fulfills many criteria for transparency and good semantic properties [11]: Moderate number of rules: fuzzy clustering helps ensuring a comprehensive sized rule base with rules that describe important regions in the data. Distinguishability: a low number of clusters leads to distinguishable rules and membership functions. Normality: by fitting parameterized ....

J. Valente de Oliveira, "Semantic constraints for membership function optimization," IEEE Transactions on Systems, Man and Cybernetics Part A: Systems and Humans, vol. 29, no. 1, pp. 128--138, 1999.

....an important aspect of these systems. Fuzzy logic helps improving the interpretability of knowledge based classifiers through its semantics that provide insight in the classifier structure and decision making process. Fuzzy logic, however, is not a guarantee for interpretability, as recognized in [2, 3]. Real effort must be made to keep the resulting rule base transparent [4, 5, 6] Two main approaches are followed in the literature: i) Select a low number of input variables in order to make a compact classifier [4, 7] and (ii) Make a large set of possible rules, by using all inputs, and then ....

J. Valente de Oliveira, "Semantic constraints for membership function optimization," IEEE Trans. FS, vol. 19, pp. 128--138, 1999.

....in fault detection, biology, medicinem etc. Fuzzy logic improves classification and decision support systems by allowing the use of overlapping class definitions and improves the interpretability of the results by providing more insight into the classifier structure and decision making process [13]. The automatic determination of fuzzy classification rules from data has been approached by several di#erent techniques: neuro fuzzy methods [6] genetic algorithm based rule selection [5] and fuzzy clustering in combination with GA optimization [12] Traditionally, algorithms to obtain ....

....Traditionally, algorithms to obtain classifiers have focused either on accuracy or interpretability. Recently some approaches to combining these properties have been reported; fuzzy clustering is proposed to derive transparent models in [9] linguistic constraints are applied to fuzzy modeling in [13] and rule extraction from neural networks is described in [8] In this paper we describe an approach that addresses both issues. Compact, accurate and linguisticly interpretable fuzzy rule based classifiers are obtained from labeled observation data in an iterative fashion. An initial model is ....

[Article contains additional citation context not shown here]

Valente de Oliveira J. (1999) Semantic constraints for membership function optimization. IEEE Trans. FS 19, 128--138.

No context found.

J. V. de Oliveira, Semantic constraints for membership function optimization, IEEE Trans. FS 19 (1999) 128--138.

No context found.

J. V. de Oliveira, Semantic constraints for membership function optimization, IEEE Trans. FS 19 (1999) 128--138.

No context found.

J. Valente de Oliveira. Semantic constraint for membership function optimization. IEEE Transactions on Systems, Man, and Cybernetics. Part A: Systems and Humans, 29(1):128--138, January 1999.

No context found.

V. de Oliveira, Semantic constraints for membership function optimization, IEEE Trans. on Systems, Man, and Cybernetics - Part A: Systems and Humans 29 (1999) 128-138.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC