Anti-Fuzzy Soft Graph Based on Neural Networks
Journal of Intelligent Decision Making and Information Science · 09.09.2026 · 3 (18s) · 3091-3106
DOI: 10.17762/jidmis.v3.6070 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 09.09.2026. Verfasst von Tahir Rasool. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This work aims to propose a comparative neural network framework using fuzzy graph, fuzzy soft graph, and anti-fuzzy soft graph, respectively, for better representation and processing of uncertain relational data. The focus is on discovering the mathematical formalism of these three mathematical structures and showing, through examples, how they can be illustrated and embedded in the learning mechanisms of a neural network. In the first place, there exists a fuzzy graph, where the weights of the vertices and the edges of the graph express the strength of the relations between connected vertices. The fuzzy soft graph is an extension of this representation that allows the inclusion of parameterized information and therefore the modelling of the relationships under various conditions. The anti-fuzzy soft graph is constructed through an explicitly given anti-fuzzy membership convention, which takes care of complementary and inverse relational information. A numerical example is worked out for each structure to show how to build the graph, compute the membership values, and check the mathematical constraints. Then a graph neural network (GNN) structure is developed to apply the membership-weighted message passing, feature aggregation, nonlinear activation, and prediction of the output. Here, vertex memberships represent node features, edge memberships control the influence of adjacent nodes, and soft parameters offer some context in feature learning. A structural representation method, uncertainty handling method, parameter flexibility, computational complexity, interpretability, and predictive performance are used as evaluation criteria to establish a systematic comparison method. If experimental data are available, the execution of a neural network can be assessed using a system of metrics such as accuracy, recall, F1-score, and loss. The proposed framework is mathematically based and offers a platform to study the differences between fuzzy graph models and their impact on neural information processing. This study provides both formal definitions and worked examples, formulates the problem as a neural network, and compares and evaluates these architectures, laying the groundwork for uncertainty-aware graph learning and, potentially, intelligent decision-making, pattern recognition, and complex network analysis.
Zitieren
APA
Rasool, T. (2026). Anti-Fuzzy Soft Graph Based on Neural Networks. Journal of Intelligent Decision Making and Information Science, 3(18s), 3091-3106. https://doi.org/10.17762/jidmis.v3.6070
Vancouver
Rasool T. Anti-Fuzzy Soft Graph Based on Neural Networks. Journal of Intelligent Decision Making and Information Science. 2026;3(18s):3091-3106. doi:10.17762/jidmis.v3.6070
BibTeX
@article{Rasool2026,
author = {Rasool, Tahir},
title = {Anti-Fuzzy Soft Graph Based on Neural Networks},
journal = {Journal of Intelligent Decision Making and Information Science},
year = {2026},
volume = {3},
number = {18s},
pages = {3091--3106},
doi = {10.17762/jidmis.v3.6070}
}
EUVIMED