Expert Systems with Applications最新文献

{"title":"A network fairness consensus model considering opinion retention utility","authors":"Dong Cheng,&nbsp;Fen Liang,&nbsp;Yong Wu","doi":"10.1016/j.eswa.2025.128848","DOIUrl":"10.1016/j.eswa.2025.128848","url":null,"abstract":"<div><div>Frequent interactions among decision-makers (DMs) in social network group decision-making (SNGDM) intensified their fairness concerns. In the consensus-reaching process of SNGDM, prior research assumes that DMs hold fairness concern for all others in social networks and only focuses on opinion compensation. However, it ignores that DMs in social networks mainly have fairness concern with those they are connected to in social-comparisons, as well as the impact of opinion reservation on the perceived fairness utility in self-comparisons. To address these two issues, this paper redefines fairness measurement in social networks and incorporates opinion retention into DMs’ fairness utility, aiming to analyze consensus fairness considering DMs’ network fairness concern in SNGDM. First, we propose the network fairness concern coefficient based on trust and opinion relationships to measure the different levels of DMs’ fairness concern for others. Then, taking into account the DM’s dual fairness concerns about opinion compensation and opinion retention, the fairness utility function is constructed based on the network fairness coefficient. Accordingly, a maximum fairness utility network consensus model is proposed. Finally, the validity of the proposed model is confirmed by the application example of enterprises’ initial carbon quota allocation. The results show that: (1) The network fairness concern coefficient enables personalized fairness assessment, and (2) Incorporating opinion retention in the fairness utility function mitigates limitations of compensation-focused fairness measures, offering a more holistic framework.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128848"},"PeriodicalIF":7.5,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572823","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A blockchain empowered federated differentiable search index framework for secure information collaboration","authors":"Qi Wang ,&nbsp;Yi Liu","doi":"10.1016/j.eswa.2025.128919","DOIUrl":"10.1016/j.eswa.2025.128919","url":null,"abstract":"<div><div>Efficient and reliable information collaboration is essential for prompt response and effective decision-making in emergency management. Current solutions face significant challenges in cross-domain data retrieval and sharing, including data fragmentation, lack of unified indexing mechanisms, and insufficient privacy protection. This paper proposes a Federated Differentiable Search Index (FeDSI) framework to address these challenges., FeDSI integrates generative retrieval, federated learning, and blockchain technology into a unified architecture to support secure, decentralized, and privacy-preserving data collaboration. The core innovation lies in adopting a differentiable search index—a learnable document indexing mechanism optimized via backpropagation—which enables semantic-based document identification directly from user queries. The model is collaboratively trained across multiple organizations using federated learning, with blockchain smart contracts ensuring transparency and verifiability of the training and retrieval processes. Experimental results on the NQ320K benchmark show that FeDSI outperforms classical sparse and dense retrieval baselines, and achieves competitive performance compared to state-of-the-art generative retrieval models, achieving a Recall@10 of 85.40 and an MRR@100 of 72.69. These findings demonstrate the effectiveness of FeDSI in supporting secure, efficient, and collaborative information retrieval in complex emergency data environments.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"296 ","pages":"Article 128919"},"PeriodicalIF":7.5,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144605662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A real-time FPGA-based radiation noise suppression and detection method with comparative analysis against deep learning techniques","authors":"Shoulong Xu ,&nbsp;Zhixiong Hou ,&nbsp;Cuiyue Wei ,&nbsp;Youjun Huang ,&nbsp;Shuliang Zou ,&nbsp;Pengfei Li ,&nbsp;Qingyang Wei","doi":"10.1016/j.eswa.2025.128906","DOIUrl":"10.1016/j.eswa.2025.128906","url":null,"abstract":"<div><div>The safe operation of nuclear facilities and the demand for real-time radiation monitoring have continued to increase. Active Pixel Sensor based nuclear radiation imaging technology has attracted significant attention due to its low power consumption and high integration capability. However, in high-radiation environments, APS devices are susceptible to interference from high-energy particle impacts, generating random high-amplitude noise that severely degrades video quality and reduces dose rate measurement accuracy. To address this issue, this paper proposes a real-time radiation noise suppression and dose detection method that combines time-domain minimum-value substitution with spatial median filtering with two-dimensional wavelet decomposition, implemented on a parallel FPGA architecture. The proposed method fully exploits the multi-stage pipelining and parallel processing capabilities of FPGAs to efficiently suppress radiation-induced noise in APS image streams and extract residual dose information at multiple scales. Experiments conducted using a ⁶⁰Co gamma-ray source on both a video test chart and a real-world scenario demonstrate that the method improves the peak signal-to-noise ratio by an average of approximately 11 dB after denoising, significantly outperforming Gaussian and low-pass filtering, and achieving comparable results to deep learning approaches such as DnCNN and Vision Transformer. Moreover, the hardware implementation does not require power-hungry GPUs, ensuring real-time performance for embedded applications. Further wavelet decomposition and pixel value fitting analyses confirm excellent linear correlation for dose rate estimation, with the Daubechies wavelet diagonal component achieving an R² as high as 0.99624. Overall, the proposed approach offers a low-power, high-efficiency engineering solution for real-time APS video denoising and dose detection in nuclear environments, providing a solid technical foundation for building FPGA-based intelligent nuclear radiation monitoring expert systems.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128906"},"PeriodicalIF":7.5,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144569853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A comprehensive review on data-level methods for imbalanced data classification","authors":"Bahareh Nikpour ,&nbsp;Farshad Rahmati ,&nbsp;Behzad Mirzaei ,&nbsp;Hossein Nezamabadi-pour","doi":"10.1016/j.eswa.2025.128920","DOIUrl":"10.1016/j.eswa.2025.128920","url":null,"abstract":"<div><div>Classification is one of the most important tasks in machine learning and data mining. Most of the classifiers are designed for data sets with equally distributed samples among the classes. Therefore, they encounter a problem with classifying imbalanced data in which one or more classes have much fewer samples than the others. Imbalanced data sets are prevalent in the real-world, so addressing this issue is of utmost importance. There have been many methods suggested to solve this problem showing promising results, a category of which is <em>data-level</em> methods being popular for their flexibility. In this paper, our goal is to review data-level methods comprehensively and categorize them from different perspectives. Also, to simplify doing future research in this field, most of the available benchmark imbalanced data sets, software, and toolboxes are introduced. Finally, existing challenges and future works are elaborated.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128920"},"PeriodicalIF":7.5,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D-PV: Enhancing PV power prediction by modeling spatial uncertainty under dynamic shading conditions","authors":"Fengze Li ,&nbsp;Dou Hong ,&nbsp;Jieming Ma ,&nbsp;Zhongbei Tian ,&nbsp;Hai-Ning Liang ,&nbsp;Jiawei Guo ,&nbsp;Kangshi Wang","doi":"10.1016/j.eswa.2025.128869","DOIUrl":"10.1016/j.eswa.2025.128869","url":null,"abstract":"<div><div>The Earth’s revolution and geographic variability introduce spatial uncertainty in photovoltaic (PV) systems. Subtle spatial variations give rise to dynamic shading conditions (DSC), which disrupt power prediction over time. Existing models often neglect to capture the effects of spatial uncertainty, and consequently struggle to address the DSC in PV systems. This paper presents a 3D-PV framework, which introduces a deblurring 3D reconstruction technique to produce spatial representations, preserving details of PV panels and their surrounding environment. Further, shadow variation matrices are constructed by the proposed ComputeShader-based shadow calculation algorithm, serving as a spatio-temporal representation to bridge the obtained spatial representations and dynamic shading variations. Building on the spatio-temporal representations, 3D-PV performs semantic fusion of shadow dynamics and irradiance signals, enabling temporally consistent power prediction under DSC. Experimental results, including ablation studies, demonstrate that precise spatial modeling effectively captures and simulates accurate shadow patterns over time. In particular, 3D-PV outperforms state-of-the-art prediction methods, achieving a 23.95 % reduction in mean squared error (MSE) for prediction accuracy. These results highlight the benefits of explicitly modeling spatial uncertainty and dynamically fusing spatio-temporal representations with irradiance signals under DSC, enabling accurate prediction of PV power.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"296 ","pages":"Article 128869"},"PeriodicalIF":7.5,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144588761","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Susceptible-infected diffusion of food safety opinion dissemination: Infrastructure-driven spread and behavior-embedded substance","authors":"Xinyue Hao ,&nbsp;Dapeng Dong ,&nbsp;Chang Liu ,&nbsp;Emrah Demir ,&nbsp;Samuel Fosso Wamba","doi":"10.1016/j.eswa.2025.128886","DOIUrl":"10.1016/j.eswa.2025.128886","url":null,"abstract":"<div><div>This study examines how food safety information disseminates across three structurally distinct Chinese social media platforms, Weibo, TikTok, and Xiaohongshu (XHS), during crisis events. Rather than serving as neutral transmission channels, these platforms are conceptualized as dynamic Information Service Systems (ISS), in which algorithmic infrastructures and content substances co-produce public meaning, emotional salience, and trust dynamics. Drawing on the Substance–Infrastructure (S-I) model, specifically Type II logic, where infrastructure drives substance, we theorize that technical mechanisms such as feed algorithms, trending systems, and visibility logics interact with semantic features like emotional tone, media modality, and narrative framing to shape the velocity, reach, and epistemic reliability of crisis communication. Employing a mixed-methods design that combines temporal Exponential Random Graph Models (ERGM), Susceptible-Infected (SI) diffusion simulations, and BERT-based sentiment analysis, we identify how different network structures, decentralized, centralized, and hybrid, interact with conformity, homophily, and neophilia to produce platform-specific information ecologies. TikTok’s architecture enables high-speed virality with minimal deliberative anchoring, limiting the platform’s ability to support trust repair; XHS facilitates high-affinity trust ecosystems led by key opinion leaders, but is vulnerable to echo chambers and insular misinformation; Weibo, with its hybrid infrastructure, supports rapid escalation and multi-directional discourse, but suffers from volatility in trust due to inconsistent epistemic control. These distinct affordances explain the asymmetric amplification of food safety narratives and the divergent trajectories of public trust, consolidation, polarization, or collapse, across platforms. As a contribution, the study introduces the Integrated Design and Operation Management (IDOM) framework, which positions platforms as reflexive control systems that must adapt to real-time signals of uncertainty and trust decay. It further underscores the need for resilient public governance that aligns institutional interventions with platform-specific logics and user cognitive baselines, advocating for a coordinated socio-technical ecosystem capable of sustaining trustworthy, inclusive, and responsive food safety communication in the digital era.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128886"},"PeriodicalIF":7.5,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multiplex graph prompt collaboration for open-set social event detection","authors":"Xiuqin Liang ,&nbsp;Jiazhen Chen ,&nbsp;Sichao Fu ,&nbsp;Wuli Wang ,&nbsp;Mingbin Feng ,&nbsp;Tony S. Wirjanto ,&nbsp;Qinmu Peng ,&nbsp;Baodi Liu ,&nbsp;Weihua Ou","doi":"10.1016/j.eswa.2025.128887","DOIUrl":"10.1016/j.eswa.2025.128887","url":null,"abstract":"<div><div>Social event detection (SED) aims to detect event types from social media messages that reflect group behavior and public concerns. Owing to the dynamic evolution nature of social media, newly occurred messages may belong to unseen event types. Recently emerged open-set SED methods introduce graph neural networks (GNN) for feature encoding and iteratively retrain model parameters via the learned pseudo-labels to adapt to new, unknown events. However, retraining the entire model with these noisy pseudo-labels inevitably causes overfitting and risk erasing information learned from known events. In this paper, we propose a novel framework to tackle the open-set SED problem by leveraging the advantages of graph prompt learning (GPL) in its fast adaptation to new data with minimal parameter modifications, termed <strong>M</strong>ultiplex <strong>G</strong>raph <strong>P</strong>rompt <strong>C</strong>ollaboration (MGPC for short). Specifically, MGPC introduces three types of graph prompts to adapt a pre-trained GNN model from old to new message graphs quickly without extensive retraining. To address the distribution shifts issue in node content, graph structures, and temporal context as new data emerges, we introduce a graph adaptation prompt and a temporal prompt, which manage information flow from old to new graphs under evolving temporal context. We further introduce a prototypical-based supervised training loss with a lightweight event embedding prompt, which facilitates quick adaptation to new event class distributions while retaining previously learned information with minimal parameter changes. These adopted prompts are fine-tuned using pseudo-labels generated according to the entropy-based uncertainty scores concerning the known classes, supplemented by an unsupervised contrastive learning component to improve inter-class discrimination for unknown events. Extensive experiments on real-world benchmarks demonstrate the effectiveness of the proposed MGPC framework in comparison to existing SED methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128887"},"PeriodicalIF":7.5,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MMT-SNN: Markovian decision and multi-threshold spike delivery integrated adaptive spiking neural network for tactile object recognition","authors":"Jing Yang ,&nbsp;Zukun Yu ,&nbsp;Changfu Zhang ,&nbsp;Shaobo Li ,&nbsp;Lin Li ,&nbsp;Zhidong Su ,&nbsp;Yixiong Feng","doi":"10.1016/j.eswa.2025.128850","DOIUrl":"10.1016/j.eswa.2025.128850","url":null,"abstract":"<div><div>Tactile object recognition represents a significant research direction within the field of robotic perception. Traditional frame-based tactile object recognition methods encounter limitations when applied to event-driven tactile data, especially under rapid dynamic change conditions. In contrast, spiking neural networks (SNNs) demonstrate higher efficiency when processing event-driven tactile data streams. However, the parameter update strategies employed by the existing SNN models typically rely on fixed learning strategies and regularization parameters, which may lead to slow convergence or entrapment in local optima when handling complex, variable tactile signals. Moreover, the current SNN models for tactile recognition often utilize single-neuron firing mechanisms, which restricts their overall neuronal expression capacities. To address these issues, we propose the MMT-SNN method, which leverages Markov decision process (MDP) principles to dynamically adjust the parameter update strategies of SNNs, thereby enhancing the accuracy and efficiency of object recognition. Additionally, a multi-threshold firing mechanism is employed to attain improved gradient propagation and increased neuronal expressiveness within the network. Experimental results demonstrate that MMT-SNN significantly outperforms the state-of-the-art approaches, achieving a 12.50% performance improvement over the classic TactileSGNet approach on the Containers-v0 dataset and a 3.61% improvement on the Objects-v0 dataset.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128850"},"PeriodicalIF":7.5,"publicationDate":"2025-07-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144580109","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Disturbed optimal power flow with renewable source and static synchronous compensator","authors":"Kaijie Xu ,&nbsp;Xiaochen Zhang ,&nbsp;Shengchen Liao ,&nbsp;Lin Qiu","doi":"10.1016/j.eswa.2025.128799","DOIUrl":"10.1016/j.eswa.2025.128799","url":null,"abstract":"<div><div>As the share of renewable energy sources increases in modern power systems, the inherent variability of these sources leads to more significant fluctuations in power load. This increased variability introduces additionalchallenges for the stability and reliability of the system. Therefore, to better model real-world power systems, this paper proposes the bus-level disturbed optimal power flow (D-OPF) problem, considering both renewable energy sources and Static Synchronous Compensators (STATCOMs). In addition, to address the uncertainties introduced by renewable energy sources and load fluctuations, this paper proposes an Enhanced Quadratic Interpolation Optimization (EQIO) algorithm. The EQIO algorithm integrates Tent chaotic mapping, Survival-of-the-Fittest selection, and dynamic opposition-based learning to improve convergence and solution accuracy under uncertain conditions. The effectiveness of the proposed EQIO algorithm is validated on the CEC2017 benchmark functions and further tested on the IEEE 30-bus and 118-bus systems under disturbed scenarios. Experimental results show that EQIO achieves Friedman Ranks of 1.1750 and 1.0733 for the 30-bus and 118-bus systems, respectively, and obtains the optimal solution in 90.08 % of all disturbed scenario tests. These outcomes demonstrate the superiority of EQIO over other algorithms in solving the D-OPF problem.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128799"},"PeriodicalIF":7.5,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144580003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"New perspectives on multivariate time series forecasting: Lightweight networks combined with multi-scale hybrid state space models","authors":"Fei Hao ,&nbsp;Junhai Qiu ,&nbsp;Xiaofeng Zhang ,&nbsp;Yepeng Liu ,&nbsp;Hua Wang ,&nbsp;Yujuan Sun ,&nbsp;Pengbin Zhang","doi":"10.1016/j.eswa.2025.128845","DOIUrl":"10.1016/j.eswa.2025.128845","url":null,"abstract":"<div><div>In the real world, applications such as industrial energy planning and urban transport planning require forecasting future trends from historical data. Due to the significance and complexity of these issues, there is an urgent need for robust prediction algorithms that can handle long-term time series forecasting. In recent years, transformer-based algorithms have emerged and demonstrated great potential. However, their computational costs are substantial, leading to inefficiency. A lightweight module called LSM is proposed to enhance the accuracy of Long-term Time Series Forecasting (LTSF). This model exhibits linear scalability and low computational costs. By effectively combining deep learning models with a hybrid state space model architecture, it efficiently captures dependencies at different scales within patches to predict global and local contexts accurately. Additionally, to further improve algorithm performance and computational efficiency, this model adopts a “strong encoder-light decoder” architecture design. Experimental results on 8 benchmark datasets demonstrate that LSM performs exceptionally well in long sequence prediction tasks by exhibiting strong robustness and effectiveness compared to State-Of-The-Art approaches (SOTA). Moreover, LSM significantly enhances accuracy while reducing computational requirements. <strong>Code availability:</strong> <span><span>https://github.com/hao-fei-hub/LSM/</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"295 ","pages":"Article 128845"},"PeriodicalIF":7.5,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572829","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}