Temporal three-way decision-making for emergency admission integrating multigranulation neighborhood rough set with Gaussian mixture-hidden Markov model

Feature selection is a core step in data analysis and is referred to as attribute reduction in rough set theory. Granular ball computing has emerged as a novel data analysis paradigm characterized by high computational efficiency, robustness, and scalability. However, in previous attribute reduction methods for interval numbers, the construction of tolerance classes and the reduction iteration process suffer from inefficiency. To address these limitations, this paper proposes an efficient attribute reduction method based on fuzzy interval-valued granular balls. This method integrates fuzzy interval-valued granular balls with an acceleration strategy based on the positive region. Specifically, we first construct tolerance classes efficiently using fuzzy interval-valued granular balls, thereby enabling a reasonable partition of the universe. We then remove redundant objects in the positive region during the reduction iteration to avoid unnecessary computations. On this basis, we further propose a conditional entropy-based algorithm for attribute reduction. Experimental results show that this algorithm substantially improves computational efficiency while maintaining high classification accuracy.

Abstract In today’s digital era, the rapid growth of deep learning technologies has made it possible to create highly realistic fake media, commonly known as deepfakes. These manipulated images, videos, and audio clips are becoming increasingly difficult to distinguish from genuine content, raising serious concerns about misinformation, privacy, and cybersecurity. While deepfakes have useful applications in areas such as entertainment and content creation, their misuse poses significant risks to individuals and society. The rapid advancement of generative artificial intelligence has enabled the creation of highly realistic synthetic media, commonly known as deepfakes, which pose significant threats to digital trust, security, and personal identity. Detecting these sophisticated forgeries requires robust machine learning models capable of identifying subtle manipulation artifacts across diverse, real-world data distributions. This paper explores the critical challenges in deepfake detection, particularly focusing on the limitations of existing single-stage and globally-averaged detection methods. We propose a comprehensive, hypothetical detection framework that integrates a local-global spatial ensemble approach with multimodal next-frame feature prediction. Furthermore, we outline a multi-faceted evaluation plan incorporating advanced metrics such as Shannon entropy and rough set theory to assess both predictive performance and model robustness. Ultimately, this paper provides a structured analysis of the practical, ethical, and technical dimensions of deploying deepfake detection systems in the wild. This research paper focuses on the use of machine learning techniques to detect deepfake content effectively. It explores different approaches, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and hybrid models that combine spatial and temporal analysis. The study also examines widely used datasets, preprocessing methods, and evaluation metrics to assess model performance. In addition, the paper highlights key challenges such as the continuous improvement of deepfake generation techniques, limited availability of diverse datasets, and difficulties in real-time detection. Finally, it discusses future directions, including the development of lightweight models and the integration of explainable artificial intelligence for better transparency. Overall, this research aims to provide a clear understanding of current deepfake detection methods and emphasize the need for more robust and adaptive solutions to address this growing threat.

In this paper, we investigate rough I - γ τ -statistical convergence of order α in neutrosophic normed spaces and establish several fundamental structural properties of the associated limit sets. We first show that the family { L j } j ≥ 0 of rough statistical limit sets is monotone with respect to the roughness parameter j , and each L j is convex under mild monotonicity assumptions on the neutrosophic components. For j = 0 , rough convergence reduces to the classical I - γ τ -statistical convergence of order α , ensuring that the limit set is a singleton. We further demonstrate that the rough limit set is always neutrosophically closed and neutrosophically convex, highlighting its stability under both topological and geometric operations. A characterization of strong I - s t γ τ α -boundedness is obtained via the non-emptiness of the rough limit set. In addition, we introduce the notion of I j − s t γ τ α -cluster points and prove that every rough limit point is a cluster point, while the cluster set remains neutrosophically closed. Finally, we show that this convergence framework unifies several classical notions of statistical and ideal convergences as particular cases.

Introduction Heart disease is a major global health problem that highlights the need for effective and accurate prediction methods. Methods This paper presents RNS-OptiDANet, a hybrid framework that combines rough set theory (RST), rough neutrosophic sets (RNS) and an optimized dual-attention neural network (OptiDANet) in order to predict heart disease. For feature selection, the QuickReduct method with the discernibility matrix (RST QRDM) was used. The features selected in RST were represented as RNS representations to deal with uncertainty in the classification process. The OptiDANet model implements Dual Attention Mechanisms such as Channel Attention (CAM) and Soft Attention Mechanism (SAM) to highlight the relevant patterns while reducing noise. The performance of the developed framework has been improved through Hyperparameter tuning using Optuna and overfitting has been avoided. Finally, classification is conducted using a Random Forest (RF) model. Results Experimental results demonstrate strong performance in terms of accuracy, precision, recall and F1-score across datasets. Discussion An eXplainable Artificial Intelligence (XAI) module is integrated to provide feature level interpretability and clinical transparency while ablation study validates the contribution of each framework component confirming the robustness and effectiveness of the proposed hybrid RNS-OptiDANet model.

Optimized Feature Selection Using Rough Sets for Intrusion Detection in 802.11 Wireless Networks

Due to the inherent subjectivity of Kansei perception, aligning the front-end styling of new energy vehicles (NEVs) with users' emotional preferences remains a complex challenge. This study proposes a data-driven framework integrating semantic mining and deep learning to quantify and optimize such emotional responses. A Latent Dirichlet Allocation (LDA) model was employed to extract four core emotional dimensions-fashion, power, technology, and sportiness-from a corpus of user-generated content (UGC). To establish a mapping between abstract emotions and concrete morphological features, rough set theory (RST) was applied for dimensionality reduction, retaining only the most influential design attributes. In addition, an attention-enhanced long short-term memory (LSTM) network optimized via a genetic algorithm (GA) was constructed to predict emotional evaluations. This hybrid model enables targeted design configuration generation for NEV front-end styling based on specific emotional indicators. The results demonstrate that the proposed approach effectively bridges the gap between qualitative user imagery and quantitative design features, providing the automotive industry with a robust emotion-oriented design support tool.

Incremental perspective for Bi-selection of instances and features by employing fuzzy rough set technique

A fuzzy deep learning ECG signal classification method via Wavelet-RL and rough set decision-making

A novel method of feature selection and information fusion for multi-source ordered information systems based on k-nearest neighbor rough sets

Some properties of two types of fuzzy rough sets on complete lattices constructed by means of overlap and grouping functions

This study improves one of the initialization methods for the k-means clustering algorithm based on a rough set neighbourhood model to ‎enhance performance in noisy datasets. The method involves data normalization, obtaining a neighbourhood threshold based on the 0.25th ‎trimmed mean of pairwise Minkowski distances, calculating cohesion and coupling degrees of the neighbourhoods and between them re-‎spectively, and obtaining the initial cluster centres as the k points having maximum cohesion degrees with minimum coupling degrees among ‎themselves. The approach was evaluated on six datasets using Silhouette, Davies–Bouldin, Calinski–Harabasz, and Dunn–Hubert indices in ‎comparison with an existing method. Results showed that the improved method outperformed the existing method on noisy datasets, achieving higher Silhouette and Dunn–Hubert scores, and lower Davies–Bouldin values, with a slight reduction in Calinski–Harabasz index in ‎one of the datasets. On the non-noisy datasets, the two methods were at par in all four performance indices. With the improved performance, showing that the improved method enhanced the stability and robustness of k-means clustering in the presence of noisy data, it can ‎be recommended for clustering noisy datasets such as gene expression, image, and signal datasets‎.

The integration of Artificial Intelligence (AI) in energy infrastructure has created a new class of specialized intermediaries for environmental control, yet their opaque decision-making poses regulatory challenges. This paper proposes a novel regulatory framework for specialized sound and vibration platform operators in the energy sector and introduces a multi-criteria decision-making (MCDM) methodology to support oversight. The methodology integrates expert neuro-behavioral data, captured via Facial Action Coding System (FACS), with a quantum picture fuzzy rough set extension and the DEMATEL (Decision-Making Trial and Evaluation Laboratory) method. The application is demonstrated through a case study of a 250 MW combined-cycle gas turbine power plant, where the goal is to select optimal noise and vibration control technologies. The analysis assesses five key technologies against compliance parameters: algorithmic transparency, data governance, system reliability, operational accountability, and consumer protection. The proposed Neuro-Quantum Picture Fuzzy Rough MCDM model achieved a forecast accuracy of 0.987 for system performance, substantially outperforming Long Short-Term Memory (LSTM (0.876)), Recurrent Neural Network (RNN (0.575)), and AutoRegressive Integrated Moving Average (ARIMA (0.551)). The primary contribution is to initiate professional dialogue on governing AI-driven energy intermediaries, balancing technological innovation with energy stability, security, and consumer welfare. The paper recommends a comprehensive regulatory framework for a new class of energy intermediaries for financial and marketing optimisation called specialised sound and vibration platform operators.

Rough graphs have emerged as a significant mathematical tool for modeling imprecise and incomplete information within graph-structured data. By leveraging rough set theory, these structures manage uncertainty through the construction of lower and upper approximations, providing a robust framework for analyzing vague or ambiguous networks. Despite the growing body of literature on rough graphs, a consistent and comprehensive theoretical foundation remains elusive. In this paper, we critically review existing partition-based approaches and identify persisting structural inconsistencies in their foundational definitions, such as the generation of invalid subgraphs and hanging edges. To address these gaps, we propose a unified, topologically consistent framework. Inspired by the theory of rough relations, we generalize vertex-based rough graphs to settings where the vertex set is granulated by a covering rather than a strict partition. Furthermore, we introduce three novel types of edge-based rough graphs derived from coverings of the edge set, demonstrating that our approach inherently preserves subgraph validity. Synthesizing these two perspectives, we propose hybrid rough graph models that establish a comprehensive multi-granulation environment for both the vertex and edge universes. Finally, we illustrate the practical utility of our proposed models through applications in social network analysis and protein-protein interaction (PPI) networks, highlighting their effectiveness in isolating core communities and identifying boundary nodes.

Access online courses, and they can be studied anywhere, anytime. The Thai government supports the use of Thai MOOC (Massive Open Online Courses) as part of Thailand’s 20-year national strategy to build competitiveness, emphasizing new ways of learning by creating the Thai MOOC Open Learning Service. This platform provides useful information to all users for accessing learning in their fields of interest. Thai MOOC offers a safe environment for individuals from diverse backgrounds to acquire valuable knowledge. This model benefits Thai MOOC by providing a system that offers a clearer picture of overall users, including the Thai Cyber University and lectures within the Thai MOOC system. Data collection and analysis employed a mixed-method approach. Qualitative research involved 21 experts in Thai MOOC, with expert consensus achieved through the rough set Delphi approach. Quantitative research gathered data from 800 actual Thai MOOC users. The analysis utilized both descriptive and inferential statistical methods. The results identified seven key factors: 1) Knowledge Sharing, 2) Thai MOOC Features, 3) Motivation and Usage, 4) Perceived Usefulness, 5) Perceived Ease of Use, 6) Attitude Toward Use, and 7) Actual Use. These factors are essential for Thai MOOC developers to understand user attitudes and improve the platform accordingly.

An analytical framework for enhancing hospital pharmacy supply chain performance using fuzzy rough set theory

Rough set based feature selection model for diabetic retinopathy classification

A novel approach to rough set model via neighborhood systems

Erratum to “Some properties of two types of fuzzy rough sets on complete lattices constructed by means of overlap and grouping functions” [Fuzzy Sets and Systems 531 (2026) 109769

Surgical robots require sub-millimeter accuracy and reliable inverse kinematics across anatomies. Population-based metaheuristics address this, but static parameters may limit achieving the needed precision for clinical use. This study introduces the Rough Sets Meta-Heuristic Schema (RSMS) for dynamic, context-aware control. RSMS categorizes agents (Elite, Boundary, Poor) via Rough Set discretization based on fitness and distribution, allocating resources accordingly without problem-specific heuristics. To demonstrate the approach’s effectiveness, RSMS was implemented within Particle Swarm Optimization and evaluated as a surgical robotics inverse kinematics solver and path planner. In simulations using three surgical problems, RS-PSO allowed upgrading of the performance of the standard PSO in terms of consistent convergence and success in tight search spaces. Statistical tests confirmed these improvements. Using a 7-DOF KUKA LBR iiwa robot and surgical benchmarks of landmark acquisition, spiral trajectory tracking, and constrained path, RS-PSO achieved success rates of 100%, 67%, and 100%, respectively, meeting surgical requirements. The results demonstrate clinical gains in accuracy, consistency, and reproducibility for minimally invasive surgery. These findings support the practical advantages of RS-PSO and, more importantly, show that the RS-MH framework can be used as a general, reusable tool to improve the robustness, precision, and reproducibility of many swarm-based meta-heuristics for surgical robotics and other applications.