As consumers face growing uncertainty about the credibility of online reviews, an important question is how they continue to evaluate service providers on digital platforms. This study examines how consumers make booking judgments on restaurant platforms when the credibility of peer-generated evaluative cues is weakened. Drawing on cue utilization theory, signaling theory, and perceived diagnosticity, we distinguish peer-generated evaluative cues (PECs) from firm-controlled operating-environment cues (FOECs). PECs refer to review-system cues generated by other consumers, whereas FOECs refer to platform cues about merchants’ operating arrangements, service attributes, and visible service environments. Two studies test the proposed conditional cue-use logic. Study 1 uses Yelp archival data to examine how PEC credibility, operational information disclosure, environmental cleanliness, and environmental orderliness relate to booking-intention-related archival responses. Study 2 uses a 2 x 2 scenario-based experiment manipulating PEC credibility and FOECs and tests the role of Perceived Diagnosticity of FOECs. The findings show that FOECs have stronger effects when PEC credibility is low, and Study 2 shows that this effect is partly explained by perceived diagnosticity. This study contributes to research on platform information cues by showing how the diagnostic value of operating-environment cues changes under weaker review credibility, and it offers practical implications for presenting clearer merchant information and environmental visual cues.

Existing studies on strategic manipulation behaviors in social network group decision-making (SNGDM) have mainly examined preference and trust separately, leaving coordinated strategic manipulation behaviors underexplored. To address this gap, we formalize preference–trust coordinated strategic manipulation behaviors (PTCSMB) as a shared-objective manipulation in which manipulators jointly manipulate preferences and trust relationships to promote a target alternative. This coordination creates structural advantages by reinforcing the channels through which manipulated preferences enter aggregation. Here, we present an integrated decision-support framework for detecting and correcting PTCSMB. The detection module extracts preference and trust evidence within observation windows, confirms target-bound trust anomalies, while the correction module restores consensus by mitigating manipulator-driven trust and preference distortions and repairing the preferences of indirectly affected decision makers. Simulations and a FilmTrust-based example first reveal the effects of PTCSMB on consensus outcomes and process stability, and further show that the proposed framework outperforms adapted baselines in preserving consensus integrity and restoring process efficiency.These findings highlight the importance of monitoring both preference and trust simultaneously in SNGDM and provide a practical approach for ensuring reliable consensus under coordinated manipulation.

Intelligent operation and maintenance (O&M) of urban rail transit signal systems (URTSS) is essential for ensuring train safety and operational efficiency. However, most O&M data exist as unstructured and sparsely labeled texts, posing major challenges for reliable knowledge extraction, semantic reasoning, and dynamic knowledge management. To address these issues, this paper proposes a unified large language model-knowledge graph framework (ULLM-KG) tailored for low-annotation, knowledge-intensive O&M environments. Firstly, a bidirectional knowledge graph construction mechanism (BKGC) is introduced to rapidly build a domain-specific initial knowledge graph. Secondly, a KG-enhanced distantly supervised entity and event extraction method (KG-DS3E) is designed to enhance critical knowledge extraction accuracy from unstructured texts. Thirdly, a prompt-driven knowledge-enhanced reasoning method (PD-KER) is proposed to improve semantic quality in fault diagnosis and maintenance recommendations. Lastly, a dynamic knowledge graph updating mechanism with temporal awareness and conflict resolution (DKG-UCF) is used to ensure efficient and accurate knowledge evolution. Based on real-world URTSS O&M data, experimental evaluations are conducted on state-of-the-art LLMs (GPT-4o, DeepSeek-V3, and Qwen3-32B). On datasets with varying annotation ratios and rare faults, ULLM-KG demonstrates significantly superior performance in knowledge extraction and reasoning tasks compared to other state-of-the-art methods. Its ability to dynamically update knowledge is also verified to be excellent. ULLM-KG provides a general solution for the intelligent O&M of URTSS under low-annotation conditions.

Next-generation smart railways require reliable real-time visual perception to ensure operational safety. However, continuously transmitting high-resolution raw video from onboard cameras to the cloud is often infeasible due to limited wireless bandwidth and fluctuating connectivity along rail lines. To address this challenge, this article presents a semantic-driven edge-cloud collaborative perception system that reduces communication overhead by transmitting task-relevant semantic masks instead of raw image frames. A lightweight semantic segmentation model deployed at the train edge extracts compact mask representations, which are further compressed and transmitted to the cloud for global risk analysis and decision support. A RailSem19-based case study demonstrates that the proposed semantic transmission mechanism significantly reduces uplink data volume and end-to-end latency under bandwidth-constrained conditions, while preserving high semantic fidelity for safety-critical railway objects. The results validate the effectiveness of communication-aware perception design for railway environments and highlight its potential for scalable deployment in next-generation intelligent transportation systems.

Short-form video (SFV)-driven e-commerce is emerging as a novel form of social commerce. However, the impact of SFV advertisements on consumer behavior remains unclear. Our study employs the Elaboration Likelihood Model (ELM) and social proof theory to explore how social proof in SFV advertisements influences consumers’ impulse buying behavior. We collected numerical data and user interaction text from Douyin and quantified real-time interaction text using grounded theory to assess the quality of social proof. The findings reveal that both the quantity and quality of social proof significantly impact impulse buying, with a moderating effect of product type on this relationship. Our study represents one of the first efforts to examine impulse buying behavior in SFV advertisements, making a significant contribution to the social commerce literature, and enhancing social proof theory. Furthermore, it offers actionable insights for sellers and marketers to effectively engage consumers and make informed, data-driven decisions.

With the escalating global energy demand, manufacturing enterprises face the challenge of enhancing energy efficiency. To address this challenge, this article centres on the flexible job shop scheduling problem considering peak power consumption constraints (FJSP-PPC) with the objective of minimizing makespan. The aim is to achieve energy-saving goals by optimizing production scheduling plans under the constraints of limited peak power. For the FJSP-PPC, three mixed integer linear programming (MILP) models are proposed based on the different ideas. In addition, an improved memetic algorithm (IMA) is designed. IMA not only adjusts the traditional active decoding scheme to adapt to the peak power limit, but also introduces the learning-based variable neighbourhood search (LVNS), which can dynamically select the application order of neighbourhood structure according to historical experience, thus the local search ability is significantly enhanced. Extensive numerical experiments are conducted to verify the superiority of the proposed models and the IMA. The experimental results show that the IMA algorithm has obvious advantages both in terms of solution quality and solution efficiency compared to existing state-of-the-art algorithms.

Cross-domain fault diagnosis (CDFD) leveraging domain generalization (DG) has attracted increasing attention due to its potential to generalize diagnostic knowledge to unseen data domains. However, the majority of current DG methods primarily model statistical dependencies, which limits their generalization performance in unobserved domains. To address this limitation, causal learning offers a natural advantage in revealing the causal linkages within diagnostic knowledge, providing a novel perspective for diagnostic guidance. Inspired by this strength, this study proposes a causality-inspired capsule network (CICN) for CDFD in the bearing field. The CICN uses a capsule network (CN) to extract features that encapsulate rich fault-related information. During the dynamic routing process, a causal gated unit is introduced to emphasize the potential causal diagnostic knowledge embedded in multisource domain fault data, enabling the model to leverage richer causal information. Furthermore, the method disentangles causal features from noncausal features and enhances the robustness of causal invariant features through causal intervention, improving the model's performance against domain interference. The effectiveness of the CICN model is validated through CDFD experiments on five datasets. The results demonstrate the superior generalization capability of the proposed approach across diverse working conditions and different machine settings.

In response to the complex treatment process and evolving medical needs of multimorbidity, multidisciplinary team (MDT) is dedicated to integrating the diagnosis opinions of experts and providing optimal treatment plans. Reaching consensus on disease treatment plans involves a dynamic and iterative group decision-making process, in which traditional methods for MDT meetings fail to address the standardized decision-making procedure, interactive trust relationships, and fuzzy information integration. Given the challenges, this study proposes a dynamic consensus framework based on dual-path feedback mechanism with q-rung orthopair fuzzy set (q-ROFS). A hybrid trust evolution model is first established within MDT, in which the trust degree is composed of inherent trust and preference similarity in each round. Then the opinion dynamics model is also introduced to the fuzzy environment. Based on trust evolution and opinion dynamics, the dual-path feedback mechanism is employed to provide references for preference adjustment and weight adjustment. Correspondingly, the calculation methods for consensus measure, preference similarity and alternative selection with q-ROFS are proposed. Additionally, a case study about vascular MDT meeting is used to illustrate the effectiveness of the proposed method. The simulation experiments are performed to verify the impact of consensus threshold, group size, individual self-confidence, and trust evolution on the proposed method. The results of the comparative analysis show that increasing the q value can expand the fuzzy information expression space while ensuring the consensus level, and the proposed method is superior to other methods in terms of more efficient and high-quality consensus results.

Failure mode and effect analysis (FMEA) is crucial for system fault detection and reliability maintenance. Despite improvements in risk assessment accuracy in recent researches, certain issues in application persist. Expert preferences are still only partially expressed, and their information aggregation process is not practical enough. Complex system failure modes have several causal coupling interactions that are rarely considered by previous research but are essential to affect risk assessment. To address these concerns, this study creates an innovative FMEA method based on the T-spherical fuzzy cognitive map (T-SFCM) that takes consensus mechanisms into consideration. First, T-spherical fuzzy sets (T-SFSs) are used to deal with the ambiguous information generated by the expert expression and provide the expert with a more flexible expression domain. Second, for the T-spherical fuzzy environment, the expert opinion aggregation procedure is enhanced. The maximum deviation method (MDM) for figuring out the weights of risk factors and the expert weight calculation method is suggested. A consensus mechanism with minimum adjustment is proposed. The expert’s hesitation in T-SFSs is considered in this procedure. Lastly, the fuzzy cognitive map (FCM) is extended to T-spherical fuzzy environment, and the T-SFCM is proposed. The causal network of fault scenarios is constructed to update the ranking results of failure modes after assessing the causal effects. A practical example illustrates the effectiveness and feasibility of the proposed method. The results indicate its superiority over other methods and its potential for application in various practical scenarios.

In actual industrial production, several operations of a job may not have precedence relationships and can be placed at any point in the process route. However, traditional flexible job shop scheduling problems (FJSP) often assume that all operations of each job must be processed in strict linear order. Therefore, this research addresses the FJSP with discrete operation sequence flexibility (FJSPDS) with the objective of minimizing the makespan. Based on existing models, two novel mixed-integer linear programming (MILP) models are formulated by improving the description methods of variables and constraints, significantly enhancing the models’ performance. Additionally, a hybrid evolutionary algorithm (HEA) is proposed to solve large-scale instances through the following three aspects. An improved encoding method is proposed, which makes the search space of the HEA and solution space of the problem more compatible and reduces the possibility of optimal solutions being missed. A special neighborhood structure is designed according to the characters of sequence-free operations, and an iterative local search method is introduced to improve the quality of the solution. A knowledge-driven reinitialization operator is developed, which generates new individuals based on the features of the historical elite population, guiding the evolution of populations, avoiding premature convergence while also avoiding falling into local optima. Finally, a total of 110 benchmark problem instances are utilized to verify the superior effectiveness of the MILP models and the HEA in solving FJSPDS.

In practical engineering scenarios, traditional fault diagnosis methods often struggle with low accuracy when dealing with imbalanced samples. As a result, various methods based on transfer learning and data augmentation have been developed to enhance fault diagnosis by capturing fault features more effectively. However, these methods tend to overlook the value of readily available real normal samples (RNSs) in fault sample generation. To leverage the features of RNSs more effectively for imbalanced fault diagnosis, we introduce a novel approach called the conditional variational auto-encoder with transfer and adversarial structures (TA-CVAE). Initially, we add perturbations from RNSs to real fault samples (RFSs) to construct the target dataset, subsequently training the VAE teacher model unsupervised to create a super fault sample (SFS) domain. Next, we blend RFSs and SFSs to train the CVAE student model, incorporating cross sequence relation knowledge distillation (CSRKD) to facilitate the transfer of knowledge from the teacher model for generating fake fault samples (FFSs). Then, RNSs, RFSs, and FFSs are utilized as inputs to refine the generator, discriminator, and classifier via adversarial learning mechanisms. Finally, we validate the performance of TA-CVAE using two rolling bearing datasets and a turnout current dataset from a real engineering scenario. Experimental results demonstrate that TA-CVAE not only produces higher quality fault samples but also achieves significantly better accuracy on imbalanced samples compared to other methods.

In order to ensure a balanced allocation of medical resources and alleviate the contradiction between the demand for medical and health services and the insufficient supply of medical resources, China continues to promote and deepen the medical reform policy, aiming to ensure the lives and health of the people. However, due to the comprehensive influence of geographical location, environment, and population, there are differences in the distribution and number of medical and health institutions, and quality and ability of doctors. The problem of uneven allocation of high-quality medical resources still exists. This study comparatively analyzes the changes in the allocation of medical resources in China over the past ten years (2011–2021), concludes the remaining deficiencies in the allocation of medical resources based on objective data, and further discusses the equity and efficiency of medical resource allocation. This research helps to meet people’s multi-level healthcare needs and promote healthy economic and social development.

Diagnostic error refers to a missed, delayed, or wrong diagnosis, which seriously affects diagnostic reliability and safety. Identifying diagnostic errors is crucial for diagnostic error research. As an important method of identifying diagnostic errors, electronic triggers (e-triggers) have several limitations, including the performance improvement bottleneck, the inability to predict diagnostic errors, and the narrow trigger range caused by inadequate data utilization. This article proposes a novel approach for identifying diagnostic errors in electronic medical records (EMRs) based on machine learning (ML) techniques. First, we design four stages for our approach, i.e., EMR utilization, ML, evaluation, and prediction. By learning the implicit identification pattern of diagnostic errors from historical EMRs, the proposed approach can find potential diagnostic error cases and predict the probability of upcoming diagnostic errors. Second, in the evaluation stage, we introduce various metrics to evaluate its performance accurately and effectively. Metrics for measuring the abilities to identify diagnostic errors and resist false negatives are introduced. Third, the study utilizes real-world EMRs of leukemia patients to illustrate and verify the proposed approach. The experimental results indicate that our approach accurately identifies most diagnostic error cases and efficiently reduces the number of false negatives, which will help protect diagnostic reliability and safety in the clinic.

The industrial application of rolling bearing fault diagnosis necessitates achieving high classification accuracy while minimizing the number of model parameters to reduce the computational resources and storage space required for the model. To meet this requirement, this study proposes a knowledge distillation convolutional neural network-deep forest (KDCNN-DF) hybrid model framework. The proposed method integrates the continuous wavelet transform (CWT) for signal data processing, a convolutional neural network (CNN) optimized by knowledge distillation (KD) for feature extraction, and a simplified multi-granular scanning (MGS) process using deep forest (DF) for fault classification. Besides, during the construction of the student models, this study found that the arrangement order of kernel sizes in the CNN convolutional layers significantly impacts the extraction of bearing fault features. Experimental validation confirmed that architecture with a smaller kernel size preceding a larger kernel size in shallow-level models is more effective. This effect is particularly pronounced after the KD process and adoption in hybrid models, resulting in higher classification accuracy. The proposed KD method reduces the parameter count of the CNN model to 5% of the original number while maintaining relatively high accuracy and significantly reducing computing time. In addition, the modeling architecture of DF has been simplified by adopting a streamlined MGS process. The proposed model achieves the highest accuracy on the original Case Western Reserve University (CWRU) datasets, with 99.75% on the 48 kHz dataset, 99.90% on the 12 kHz dataset, and a perfect 100% on the Ottawa dataset. These results surpass the accuracy of existing methods.

Semantic textual similarity (STS) is a fundamental task in the field of natural language processing (NLP). Recent advances demonstrate that deep-learning-based approaches can achieve excitingly accurate STS measurement. However, existing studies cannot capture the spatial location of important information by attention mechanisms, fail to model sentences from the perspective of overall sentences, and neglect to deal with semantic fuzziness. In this article, we propose a novel double attentive fuzzy convolutional neural network (DAFCNN) to measure STS more accurately with the consideration of semantic fuzziness. This article first introduces the spatial attention module and combines it with the improved attentive convolutions to create a multigranularity convolutional neural network in DAFCNN, which not only extracts critical spatial location information but also models sentences from multiple perspectives at word and sentence levels. Second, DAFCNN pioneers a fuzzy learning module (FLM) to fulfill the extraction of fuzzy semantic features. By using the fuzzy membership function, fuzzy aggregation operator, and trainable parameters and weights, FLM can map sentence representations to fuzzy space to constitute representations with more accurate and rich semantics. Third, compared with various state-of-the-art STS models, DAFCNN decreases by 14.57% mean-square error, increases by 4.61% Pearson's gamma and 8.57% Spearman's rho on STS score datasets, and increases by 3.39% accuracy and 2.41% F1-score on semantic classification dataset. The ablation experiment demonstrates the effectiveness of each module of DAFCNN. Finally, the experimental results also indicate that FLM is a promising new attempt to incorporate fuzzy set theory in the NLP field.

Medical Knowledge Graphs (MKGs) are vital in propelling big data technologies in healthcare and facilitating the realization of medical intelligence. However, large-scale MKGs often exhibit characteristics of data sparsity and missing facts. Following the latest advances, knowledge embedding addresses these problems by performing knowledge graph completion. Most knowledge embedding algorithms rely solely on triplet structural information, overlooking the rich information hidden within entity property sets, leading to bottlenecks in performance enhancement when dealing with the intricate relations of MKGs. Inspired by the semantic sensitivity and explicit type constraints unique to the medical domain, we propose BioBERT-based graph embedding model. This model represents an evolvable framework that integrates graph embedding, language embedding, and type information, thereby optimizing the utility of MKGs. Our study utilizes not only WordNet as a benchmark dataset but also incorporates MedicalKG to compare and corroborate the specificity of medical knowledge. Experimental results on these datasets indicate that the proposed fusion framework achieves state-of-art (SOTA) performance compared to other baselines. We believe that this incremental improvement provides promising insights for future medical knowledge graph completion endeavors.

Individuals with bounded rationality may make incorrect decisions regarding innovation projects. Organizational decision mechanisms aim to minimize the impact of individual fallibility and integrate collective wisdom. In recent years, consensus mechanism has posed a challenge to the traditional hierarchy, but its superiority in different decision scenarios is still unknown. We created decision scenarios based on positions (middle manager or grassroots employee) of the proposer of the innovation project, market volatility, and decision timeliness. Simulation results using Agent Based Modeling indicate the following: If the proposer is a middle manager and decision timeliness is unimportant, consensus in low-density networks is preferred; otherwise, hierarchy is preferred. If the proposer is a grassroots employee, the best-performing mechanism, in order, is consensus in low-density networks, consensus in high-density networks, and hierarchy as the discount interest rate increases. Market volatility drives the advance of the two critical points of the discount interest rate. Furthermore, we propose a hybrid organizational decision mechanism that can balance authority and consensus. It combines the high decision accuracy of consensus with the high decision efficiency of hierarchy, addressing the long decision time in consensus and the information asymmetry and information loss in hierarchy. Simulation results show that the proposed hybrid decision mechanism is only slightly inferior to consensus mechanism when market volatility is very low or decision timeliness is unimportant.

Multi-attribute group decision-making (MAGDM) is one of the research hotspots in human cognitive and decision-making theory. However, there are still challenges to the existing MAGDM methods in modeling uncertain linguistics of decision-makers’ (DMs’) cognitive information and objectively obtaining weights. Therefore, this paper aims to develop a new MAGDM method considering incomplete known weight information under spherical uncertain linguistic sets (SULSs) to model uncertain information in MAGDM problems. The method mainly includes the following aspects. Firstly, a new concept, which enables an intuitive evaluation of neutral membership and hesitancy degrees at the linguistic evaluation, has been is first developed for capturing the more uncertain information. Secondly, the cosine similarity measure (CSM) and cross-entropy measure (CEM) are widely used to measure ambiguous information because of their robustness of measurement results. The CSM and CEM are extended to SULSs to calculate the DMs’ and attributes weights quantitively, respectively. Thirdly, in terms of effective integration of fuzzy information to obtain more accurate decision results, the Hamy mean (HM) and dual Hamy mean (DHM) operators are valued due to their consideration of the interrelationships between inputs. Two extension operators, named spherical fuzzy uncertain linguistic weight HM and DHM, are proposed to integrate spherical fuzzy uncertain linguistic information in the third stage. In the experiment, a decision case is presented to illustrate the applicability of the proposed method, and results show the effectiveness, flexibility and advantages of the proposed method are demonstrated by numerical examples and comparative analysis.

Fuzzy c-means (FCM) algorithm is an unsupervised clustering algorithm that effectively expresses complex real world information by integrating fuzzy parameters. Due to its simplicity and operability, it is widely used in multiple fields such as image segmentation, text categorization, pattern recognition and others. The intuitionistic fuzzy c-means (IFCM) clustering has been proven to exhibit better performance than FCM due to further capturing uncertain information in the dataset. However, the IFCM algorithm has limitations such as the random initialization of cluster centers and the unrestricted influence of all samples on all cluster centers. Therefore, a novel algorithm named equidistance index IFCM (EI-IFCM) is proposed for improving shortcomings of the IFCM. Firstly, the EI-IFCM can commence its learning process from more superior initial clustering centers. The EI-IFCM algorithm organizes the initial cluster centers based on the contribution of local density information from the data samples. Secondly, the membership degree boundary is assigned for the data samples satisfying the equidistance index to avoid the unrestricted influence of all samples on all cluster centers in the clustering process. Finally, the performance of the proposed EI-IFCM is numerically validated using UCI datasets which contain data from healthcare, plant, animal, and geography. The experimental results indicate that the proposed algorithm is competitive and suitable for fields such as plant clustering, medical classification, image differentiation and others. The experimental results also indicate that the proposed algorithm is surpassing in terms of iteration and precision in the mentioned fields by comparison with other efficient clustering algorithms.

Multivariate time series classification (MTSC) is a crucial machine learning problem prevalent across various real-life domains. Traditional deep learning approaches with high accuracy in MTSC are often criticized for their “black box” nature, offering no insight into their operational mechanisms or decision-making processes. Explainable artificial intelligence (XAI) becomes a key idea for dealing with such limitations in decision-sensitive areas. While some researchers have explored the interpretability of MTSC, the majority have focused on elucidating the relationship between variables over time, neglecting the intricate connections among different variables. To address the interpretability issue in MTSC, we introduce an explainable dual-mode convolutional neural network (XDM-CNN) designed specifically for MTSC. The proposed XDM-CNN framework comprises two modules: a classification module and an explanation module. The classification module can ensure exceptional classification performance by combining both one-dimensional and two-dimensional convolutional neural networks, while the explanation module can mine the underlying logical relationships among variables during the classification process by utilizing a synergy of visualization and quantification techniques. We validate our approach using 26 datasets sourced from the University of East Anglia (UEA) archive. The experimental results demonstrate that XDM-CNN not only exhibits excellent performance in classification accuracy, but also has strong explainability. By combining visual and numerical explanation, the hidden logical relationships among multivariate time series are explored and interpreted, providing human users with a decidable basis for classification decisions.