Membership function

About: Membership function is a research topic. Over the lifetime, 15795 publications have been published within this topic receiving 418366 citations.

An algorithmic approach for fuzzy inference

Abstract: To apply fuzzy logic, two major tasks need to be performed: the derivation of production rules and the determination of membership functions. These tasks are often difficult and time consuming. This paper presents an algorithmic method for generating membership functions and fuzzy production rules; the method includes an entropy minimization for screening analog values. Membership functions are derived by partitioning the variables into the desired number of fuzzy terms and production rules are obtained from minimum entropy clustering decisions. In the rule derivation process, rule weights are also calculated. This algorithmic approach alleviates many problems in the application of fuzzy logic to binary classification.

Fuzzy Rough Sets and Multiple-Premise Gradual Decision Rules

Abstract: We propose a new fuzzy rough set approach which, differently from most known fuzzy set extensions of rough set theory, does not use any fuzzy logical connectives (t-norm, t-conorm, fuzzy implication). As there is no rationale for a particular choice of these connectives, avoiding this choice permits to reduce the part of arbitrary in the fuzzy rough approximation. Another advantage of the new approach is that it is based on the ordinal properties of fuzzy membership degrees only. The concepts of fuzzy lower and upper approximations are thus proposed, creating a base for induction of fuzzy decision rules having syntax and semantics of gradual rules. The proposed approach to rule induction is also interesting from the viewpoint of philosophy supporting data mining and knowledge discovery, because it is concordant with the method of concomitant variations by John Stuart Mill. The decision rules are induced from lower and upper approximations defined for positive and negative relationships between credibility degrees of multiple premises, on one hand, and conclusion, on the other hand.

A framework for modeling human-like driving behaviors for autonomous vehicles in driving simulators

Abstract: A framework for modeling driver behavior within driving simulators is described in this paper. This framework serves as a basis for building human- like driving behavior models for autonomous vehicles operating within the virtual environment of a driving simulator. The framework consists of four units, the Perception Unit, the Emotions Unit, the Decision- making Unit (DMU), and the Decision- implementation Unit (DIU). The Perception Unit defines how the model perceives its environment in local and global terms. The Emotions Unit defines how the model responds emotionally to its environment. The DMU investigates the environment for possible actions that might potentially serve the model's emotional demands. And finally the DIU tries to implement these decisions when a traffic condition, perceived as safe enough for such an implementation, emerges. Each of these units has its own set of fuzzy variables and fuzzy ifthen rules. Any driving model, that is based on this framework, should provide membership function parameters for these fuzzy variables in accordance with the category of human driving behavior this model is targeting. Our framework addresses decision making and implementation at the maneuvering and operational levels of the driving task. Decisions at the planning level are addressed through a script- based traffic controller. The present model is limited to simulating human behaviors when driving in a two- lane rural environment.

Data-Core-Based Fuzzy Min–Max Neural Network for Pattern Classification

Abstract: A fuzzy min-max neural network based on data core (DCFMN) is proposed for pattern classification. A new membership function for classifying the neuron of DCFMN is defined in which the noise, the geometric center of the hyperbox, and the data core are considered. Instead of using the contraction process of the FMNN described by Simpson, a kind of overlapped neuron with new membership function based on the data core is proposed and added to neural network to represent the overlapping area of hyperboxes belonging to different classes. Furthermore, some algorithms of online learning and classification are presented according to the structure of DCFMN. DCFMN has strong robustness and high accuracy in classification taking onto account the effect of data core and noise. The performance of DCFMN is checked by some benchmark datasets and compared with some traditional fuzzy neural networks, such as the fuzzy min-max neural network (FMNN), the general FMNN, and the FMNN with compensatory neuron. Finally the pattern classification of a pipeline is evaluated using DCFMN and other classifiers. All the results indicate that the performance of DCFMN is excellent.