International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems

Grouping functions, as one new case of not necessarily associative particular binary aggregation functions, have been proposed in the literature for their vast applications in fuzzy community detection problems, image processing and decision making. On the other hand, due to the wide applications in fuzzy reasoning, fuzzy control and approximate reasoning, the investigations of fuzzy implications derived from specific binary aggregation functions become a natural and hot research topic. In this paper, we focus on this research area and consider the ${\mathcal{ℐ}}_{\mathcal{𝒢},\mathcal{𝒩}}$-implications induced from quasi-grouping functions and negations on bounded lattices. To be precise, firstly, by removing the continuity condition, we extend the notion of grouping functions on the unit closed interval to the so-called quasi-grouping functions on bounded lattices. Secondly, we give some basic properties and two construction methods of quasi-grouping functions on bounded lattices. Finally, we give the concept of ${\mathcal{ℐ}}_{\mathcal{𝒢},\mathcal{𝒩}}$-implications and focus on the conditions under which they can satisfy the certain algebraic properties possessed by implications on bounded lattices.

This paper investigates uninorms that are neither conjunctive nor disjunctive on bounded lattices. New methods are introduced for construction of such uninorms, where some restrictions on the identity and the annihilator are considered. In particular, new types of idempotent uninorms on bounded lattices are obtained. Furthermore, some specific examples are provided to illustrate that these constructions differ from the existing ones.

In this paper, we extend the various ideas of convergence such as convergence in distribution, in mean, and in measure along with the convergence uniformly almost surely of the uncertain fuzzy sequence of real numbers in statistical context. We also establish some interesting results among these convergence.

Decisions in real-life can be adversely affected by various uncertainty-sources such as perception-diversity, data-structure and analytical tools. Fuzzy clustering can successfully handle the uncertainties while recognizing patterns in any given data. Nevertheless, type-1 fuzzy clustering techniques has uncertainties on account of precise-nature of primary memberships. Type-2 fuzzy clustering are preferred by many researchers to manage uncertainty in its type-1 version. In type-2 fuzzy clustering, order of fuzziness (fuzzifier) is obtained by interval-valued or general type-2 fuzzy sets. Interval type-2 fuzzy clustering can reduce the computational complexity of type-2 fuzzy set mathematics. However, general type-2 fuzzy clustering scrutinizes uncertainty in fuzzifier using linguistic sets. Interval and general type-2 fuzzy clustering algorithms include type-reduction approaches to obtain type-1 fuzzy sets. Besides, full type-2 fuzzy c-means can be used as a foundation approach in type-2 fuzzy inferences. Although this algorithm includes precise-fuzzifier, it gives a point of view to practically calculate secondary memberships. In this paper, an adaptive type-2 fuzzy clustering algorithm is proposed to manage the uncertainty-sources with a self-reduction procedure. Several numerical results and comparisons are given to demonstrate the achievement of this proposed algorithm. The performance of the proposed algorithm is compared with type-1 and type-2 versions for various multi-dimensional pattern sets from UCI-patterns, Berkeley segmentation database and a real-life application related to sustainable supplier selection in an automotive industry. Consequently, the proposed algorithm reveals fast, convenient and precise results.

Game Theory (GT) is the study of strategic decision making. By virtue of its importance, several GT based methodologies for Feature Selection (FS) are proposed in recent times. FS problem can be abstracted as a game by considering each feature as a player and their values as their strategies. Additionally, overall goal of the game is set to classify a data instance appropriately. Most of the existing GT based FS techniques are restricted to Zero Sum Games, Non-Zero Sum Games and Cooperative Games. The classical setting of assuming that all the details of players are known to all players cannot hold in many real-world problems. When the given features are independent, they cannot be treated alike and a characteristics based uncertainty persists among the features. This uncertainty is handled by none of the game forms used in the existing methods. Unlike the mentioned game techniques, Bayesian Games (BG) address the games with imperfect information. This paper investigate the FS problem in terms of BG and proposes a novel method to select the best features. The proposed BG based FS method is a filter type FS method and it starts with identifying Principle Features (PF) and proceeds to play global pairwise Bayesian games between those PF to obtain feature scores. Later, the features are ranked using these scores. In the final stage, a forward selection method with Support Vector Machine (SVM) is used to evaluate the classification performance of the ranked features and helps in the selection of the optimal set of features. Besides these, to improve the scalability of the proposed method MapReduce paradigm is exploited. In order to show the efficacy of the proposed method, experiments are carried out with seven real-world datasets from UCI and Statlog repositories. The promising results showed a significant improvement in the classification performance with fewer selected features than which is achieved using the existing methods.

Time Series (TS) models are well-known techniques that help to predict the weather in a certain time period. The traditional TS prediction models take more prediction time, overfitting and under-fitting of training data. In addition, state-of-art method like regression automata technique’s computational complexity is high due to the learning process based on heuristic method. In this study, learning process of the Cellular Automata (CA) is performed by the Support Vector Regression (SVR) method. The better discretize property of the SVR increases the learning rate and reduces the complexity of the CA. The proposed model is named as a Range Regression Automata Model based on Support Vector (RRAMSV) and this model calculates the multiple temperature attributes like soil temperature, air temperature, relative humidity and soil humidity. The proposed RRAMSV method has higher performance in most of the time-series prediction compared to Range Linear Regression (RLR), Range Logistic Regression (RLOR) and Convolutional Neural Network (CNN) — Long Short Term Memory (LSTM). The proposed RRAMSV method has Mean Absolute Error (MAE) of 0.06 and existing CNN-LSTM method has 1.65 MAE.

Associated path detection is considered as the major concern of the traditional shortest path issue. The associated path is generally represented by the shortest distance among the source and destination. In the transportation network, distance or cost detection may identify this associated path. Specifically, it is very important to discover the shortest distance that has a minimum number of nodes, and it will give the most optimized result. In this paper, the Fuzzy based Pareto Optimal (FPO) approach is used to discover the shortest paths in a network graph. Initially, the FPO technique finds the shortest paths in a network by using set of rules. Then, the Lexicographical model uses a set of rules to rank the shortest distance based on minimum distance value. From the ranking results, the optimal shortest path is selected based on the proposed Ant Lion Optimization (ALO) algorithm. So, this paper achieves multi objectives like shortest path ranking and selection of the optimal shortest path. Time, distance or cost, convergence time, fitness function, and mean square error are the parameters used to relate the performance of the proposed technique with state-of-the-art techniques. Comparative results display the robustness and proficiency of the proposed system with several works.

Please send your abstracts (or copies of papers that you want to see reviewed here) to vladik@utep.edu, or by regular mail to Vladik Kreinovich, Department of Computer Science, University of Texas at El Paso, El Paso, TX 79968, USA…