Applied Soft Computing最新文献

Transformer spectral optimization: From gradient frequency analysis to adaptive spectral integration 变压器频谱优化：从梯度频率分析到自适应频谱积分

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-25 DOI: 10.1016/j.asoc.2025.113637

Zhigao Huang, Musheng Chen, Shiyan Zheng

{"title":"Transformer spectral optimization: From gradient frequency analysis to adaptive spectral integration","authors":"Zhigao Huang, Musheng Chen, Shiyan Zheng","doi":"10.1016/j.asoc.2025.113637","DOIUrl":"10.1016/j.asoc.2025.113637","url":null,"abstract":"<div><div>This paper explores a novel perspective on Transformer optimization by analyzing gradient characteristics in the frequency domain. First, we systematically quantify spectral differences between attention and MLP layer gradients, revealing that attention gradients consistently exhibit higher frequency content (23% higher mean frequency, 37% more prominent high-frequency components) compared to MLP gradients. Second, we demonstrate the potential of using spectral features for monitoring training dynamics, finding a strong correlation (r=-0.82) between early-stage spectral entropy and final validation loss. Third, building on these insights, we introduce Adaptive Spectral Integration (ASI), an optimization framework that selectively filters gradient spectra during training. Our experiments on GPT2-small with standard datasets (Penn Treebank and WikiText-2) show that ASI achieves notable inference speed improvements (6.3%-9.1%) and training time reductions (13.2%-18.8%) while maintaining comparable model quality. However, cross-architecture validation with BERT-style models reveals that ASI’s efficiency benefits are architecture-dependent, showing limited improvements on bidirectional models. These findings provide evidence that frequency domain analysis offers valuable insights for optimizing autoregressive Transformer models, while highlighting the need for architecture-aware spectral optimization strategies.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113637"},"PeriodicalIF":7.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711626","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}

引用次数: 0

Federated learning with three-way decisions for privacy-preserving multicloud resource scheduling 具有三向决策的联邦学习，用于保护隐私的多云资源调度

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-25 DOI: 10.1016/j.asoc.2025.113634

Chunmao Jiang, Lirun Su

{"title":"Federated learning with three-way decisions for privacy-preserving multicloud resource scheduling","authors":"Chunmao Jiang, Lirun Su","doi":"10.1016/j.asoc.2025.113634","DOIUrl":"10.1016/j.asoc.2025.113634","url":null,"abstract":"<div><div>This paper introduces the Federated Three-Way Decision System (F3WDS), a novel framework for multicloud resource scheduling that integrates federated learning with the three-way decision theory to address the challenges of resource heterogeneity, decision uncertainty, and data privacy. By combining privacy-preserving collaborative learning with nuanced decision-making (positive, boundary, and negative regions), the F3WDS optimizes resource allocation across multiple cloud providers while adhering to strict data sovereignty requirements. We provide rigorous theoretical guarantees, including convergence analysis, privacy bounds, and performance bounds, to demonstrate the reliability of the system. Extensive experiments on synthetic and real-world datasets demonstrate that F3WDS achieves significant improvements over state-of-the-art methods: 5%–14% higher resource utilization, 60% lower privacy loss, and 30% reduced cross-cloud latency. The framework’s scalability, robustness to stragglers, and favorable privacy-utility trade-off make it a solution for privacy-sensitive multicloud environments, with implications for future research on distributed computing and privacy-aware resource management.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113634"},"PeriodicalIF":7.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711627","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}

引用次数: 0

Measuring building information modeling user satisfaction by using active interpretable machine learning 通过使用主动可解释的机器学习来测量建筑信息建模用户满意度

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-25 DOI: 10.1016/j.asoc.2025.113663

Wei-Chih Wang, Shyn-Chang Huang, Hsu-Pin Wang, Minh-Tu Cao

{"title":"Measuring building information modeling user satisfaction by using active interpretable machine learning","authors":"Wei-Chih Wang, Shyn-Chang Huang, Hsu-Pin Wang, Minh-Tu Cao","doi":"10.1016/j.asoc.2025.113663","DOIUrl":"10.1016/j.asoc.2025.113663","url":null,"abstract":"<div><div>Accurately predicting building information modeling (BIM) user satisfaction (US) is essential for proactively addressing implementation challenges, ensuring effective adoption, and maximizing return on investment in BIM technologies in construction projects. Accordingly, this study developed advanced, interpretable boosting ensemble models to predict BIM US by integrating the forensic-based investigation (FBI) algorithm with gradient boosting machine, light gradient boosting machine, adaptive boosting (AdaBoost), extreme gradient boosting, and random forest algorithms. To validate the proposed models and establish a dataset, a comprehensive survey was conducted on 70 construction projects in Taiwan that used BIM technologies to support design work. Subsequently, the synthetic minority oversampling technique (SMOTE) was integrated into the proposed models to address the data imbalance problem. The results indicated that among all models, the FBI-AdaBoost-SMOTE model exhibited the highest performance, achieving accuracy, precision, recall, and F1 scores of 88.6 %, 90.6 %, 88.6 %, and 87.8 %, respectively. The FBI-AdaBoost model based on Shapley additive explanations identified contextual analysis and visualization, project scale, and cost estimates as key determinants of BIM US. Overall, this study presents an advanced machine learning framework for predicting BIM US and identifying key influencing factors for BIM US. It also provides actionable insights for stakeholders to enhance BIM implementation and user experience. In addition, this study highlights the potential of predictive modeling for optimizing the adoption of BIM in the architecture, engineering, and construction industry.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113663"},"PeriodicalIF":7.2,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711630","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}

引用次数: 0

Multi-source manifold space domain adaptation with a full-thresholding residual network for machinery fault diagnosis 基于全阈值残差网络的多源流形空间域自适应机械故障诊断

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-24 DOI: 10.1016/j.asoc.2025.113611

Wenhua Chen, Jianbin Li, Sixing Wu

{"title":"Multi-source manifold space domain adaptation with a full-thresholding residual network for machinery fault diagnosis","authors":"Wenhua Chen, Jianbin Li, Sixing Wu","doi":"10.1016/j.asoc.2025.113611","DOIUrl":"10.1016/j.asoc.2025.113611","url":null,"abstract":"<div><div>Unsupervised multi-source domain adaptation, which has been intensively investigated in recent years, is promising in handling fault diagnosis tasks when no labeling is available on the target datasets. Most approaches aim to learn the domain-invariant features of all domains in common feature spaces. However, addressing practical scenarios in which data comes from multiple domains with large shifts remains challenging. Hence, a new multi-source manifold space domain adaptation method (MMSDA) with a full-thresholding residual network is proposed for machinery fault diagnosis, in which specific domain-invariant features of the source and target domains are learned. First, a full-thresholding residual convolutional neural network (FTRCNN) is designed to extract useful features from both source and target domains, which are then projected into a specific domain feature space. Then, the proposed manifold neighbor consistency (MNC) domain alignment algorithm maps the feature space to a manifold space, ensuring that the samples maintain local neighbor geometric relations. Additionally, multi-kernel maximum mean discrepancy is used to reduce the inter-domain differences. Thus, the specific domain-invariant features of each source and target domain pair in the manifold feature space are extracted. Finally, the domain-specific classifier consistency (DSCC) loss is designed to minimize the shifts in all classifiers. Through experiments on three benchmarks, the proposed method demonstrates promising results on popular rotating machinery datasets for fault diagnosis.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113611"},"PeriodicalIF":7.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711628","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}

引用次数: 0

Adversarial purification using random encoding networks 使用随机编码网络的对抗性净化

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-24 DOI: 10.1016/j.asoc.2025.113604

Yuxin Gong, Shen Wang, Xunzhi Jiang, Tingyue Yu, Fanghui Sun

{"title":"Adversarial purification using random encoding networks","authors":"Yuxin Gong, Shen Wang, Xunzhi Jiang, Tingyue Yu, Fanghui Sun","doi":"10.1016/j.asoc.2025.113604","DOIUrl":"10.1016/j.asoc.2025.113604","url":null,"abstract":"<div><div>Deep neural networks (DNNs) have revealed vulnerabilities to adversarial examples, which can deceive models with high confidence. This has given rise to serious threats in security-critical domains. Adversarial defense methods have been extensively studied to counter adversarial attacks. Adversarial purification, as a major defense strategy, attempts to recover adversarial examples to clean counterparts by filtering out perturbations. However, many purification defenses struggle against white-box attacks where the target and defense models are known. Additionally, the training processes against specific attacks can compromise models’ adaptability to unknown attacks, and purification operations may destroy key features of inputs. In this paper, we propose the random encoding network (REN), which consists of a random encoding denoiser and a diverse classifier to enhance the robustness of adversarial purification defense models. The internal part of the denoiser leverages adversarial sparse coding to purify examples by filtering out perturbations and noise as much as possible while preserving critical features of inputs. The external part of the denoiser employs a dynamic random mechanism to implement random encoding, thereby enhancing the models’ uncertainty. Moreover, the classifier is subjected to a diversity constraint to promote variation among random sub-models. Experimental results demonstrate that REN exhibits strong defensive generalization capabilities, effectively countering adversarial examples across diverse attack types and settings. For the CIFAR-10 and SVHN datasets, the clean-trained REN achieves average adversarial accuracies of 63.26% and 59.78% against white-box attacks, while the adversarial-trained REN achieves 68.27% and 72.39%, respectively. When faced with unknown attack scenarios, REN is more effective than state-of-the-art defense methods.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113604"},"PeriodicalIF":7.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703052","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}

引用次数: 0

Overlapping community detection based on graph attention autoencoder and self-trained clustering 基于图注意自编码器和自训练聚类的重叠社区检测

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-24 DOI: 10.1016/j.asoc.2025.113584

Weitong Zhang , Wenxu Wang , Ronghua Shang , Songhua Xu

{"title":"Overlapping community detection based on graph attention autoencoder and self-trained clustering","authors":"Weitong Zhang , Wenxu Wang , Ronghua Shang , Songhua Xu","doi":"10.1016/j.asoc.2025.113584","DOIUrl":"10.1016/j.asoc.2025.113584","url":null,"abstract":"<div><div>Existing methods for detecting overlapping communities often rely solely on the attributes of the nodes and the network structure, but fail to make full use of the similarity relationship between nodes and their neighbors. Additionally, these methods lack effective utilization of a priori information, making it challenging to extract information about community structure and nonlinear data information in overlapping communities. To address these issues, a method for detecting overlapping communities based on a graph attention autoencoder and self-training clustering (GASTC) is proposed. Firstly, GASTC utilizes the graph attention autoencoder for overlapping community detection. The fuzzy modularity maximization method is embedded into the graph attention autoencoder to perform soft allocation of network nodes. Simultaneously, targeted learning is conducted based on the weights assigned to nodes and their neighboring nodes to capture the interactions between overlapping nodes and different communities. GASTC also designed a structural similarity function suitable for detecting overlapping communities. The community structure within overlapping communities is extracted through a semi-supervised learning approach that not only utilizes label information to enhance the prior, but also introduces connection probabilities between nodes. This enables the calculation of the structural similarity between the known network structure and unlabeled nodes. Finally, subspace clustering is used for self-training, where the cluster labels is used to supervise the learning of potential node features and self-expression coefficient matrices. The obtained self-expression coefficient matrix is used to guide the division of clusters, to capture the non-linear data information in overlapping communities. Experimental results on six datasets demonstrate that GASTC can achieve higher accuracy in overlapping community detection tasks, especially in networks with more complex structures.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113584"},"PeriodicalIF":7.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711625","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}

引用次数: 0

Dynamic Self-Attention Network based opinion formation over dynamic social networks with application to live-streaming 基于动态自关注网络的动态社交网络意见形成及其在直播中的应用

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-24 DOI: 10.1016/j.asoc.2025.113598

Haixia Mao , Yiyi Zhao , Min Xu , Jianglin Dong , Jiangping Hu

{"title":"Dynamic Self-Attention Network based opinion formation over dynamic social networks with application to live-streaming","authors":"Haixia Mao , Yiyi Zhao , Min Xu , Jianglin Dong , Jiangping Hu","doi":"10.1016/j.asoc.2025.113598","DOIUrl":"10.1016/j.asoc.2025.113598","url":null,"abstract":"<div><div>Opinion Dynamics (OD) is a widely used framework for studying the evolution of group opinions in complex social networks. However, most existing models primarily focus on static networks or small-scale dynamic networks, leaving large-scale dynamic networks largely underexplored. To address this gap, this paper investigates the evolution of group opinions and behavior prediction in large-scale dynamic social networks, using live-streaming data as a case study. Specifically, we propose two novel models: (1) a Dynamic Community Network (DCN) model, which constructs dynamic networks based on real-world big data, and (2) the Real-Time Dynamic Self-Attention Network Hegselmann–Krause (RT-DySAT-HK) model, which integrates Dynamic Graph Neural Networks (DGNNs) with OD to model the evolution of group opinions and predict behaviors. Through empirical analysis and simulations, we demonstrate that user behaviors in dynamic live-streaming networks are significantly influenced by community stability. Notably, during the early and middle stages of live-streaming, community size plays a critical role in attracting and retaining users. Moreover, the RT-DySAT-HK model proves highly effective in real-time group behavior prediction, particularly in large-scale dynamic networks. Compared to baseline models, it excels in extracting high-quality node representations and achieving accurate behavior predictions. Additionally, our findings reveal that the evolution of group opinions is influenced by multiple factors, including the contradictory effects of opinion weights and update speeds, which can lead to opinion polarization. Excessively slow update speeds may also result in opinion fragmentation. These insights contribute to a deeper understanding of OD in large-scale, dynamic environments and offer practical implications for predicting and managing group behaviors in real-world scenarios.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113598"},"PeriodicalIF":7.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703046","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}

引用次数: 0

Artificial intelligence-based semi-supervised crop and weed semantic segmentation 基于人工智能的半监督作物和杂草语义分割

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-24 DOI: 10.1016/j.asoc.2025.113662

Chaeyeong Yun, Yu Hwan Kim, Sung Jae Lee, Su Jin Im, Kang Ryoung Park

{"title":"Artificial intelligence-based semi-supervised crop and weed semantic segmentation","authors":"Chaeyeong Yun, Yu Hwan Kim, Sung Jae Lee, Su Jin Im, Kang Ryoung Park","doi":"10.1016/j.asoc.2025.113662","DOIUrl":"10.1016/j.asoc.2025.113662","url":null,"abstract":"<div><div>Accurate segmentation of crop and weed by farming robot camera can increase crop production and reduce unnecessary herbicide, which is a fundamental task in the field of sustainable and precision agriculture. However, obtaining the pixel-wise annotation of training data manually is expensive. As a solution to address this limitation, semi-supervised learning leverages a small amount of labeled data and a large amount of unlabeled data for learning. In this context, we propose a network based on vector quantization and prototype loss for semi-supervised crop and weed semantic segmentation (VQP-Net). VQP-Net achieves a strong performance in terms of consistency regularization through the implementation of a vector quantization module and prototype loss, and is capable of extracting discriminative features of crops and weeds, which are often indistinguishable. We conducted experiments using the proposed method with three open datasets: BoniRob, crop/weed field image, and rice seedling and weed datasets. The crop and weed segmentation accuracies based on mean intersection over union (<em>mIOU</em>) for the three datasets were 0.8643, 0.8329, and 0.7623, respectively, demonstrating that this method outperformed the state-of-the-art methods.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113662"},"PeriodicalIF":7.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711435","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}

引用次数: 0

PortfolioZero: A stock portfolio model based on deep reinforcement learning PortfolioZero：基于深度强化学习的股票投资组合模型

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-23 DOI: 10.1016/j.asoc.2025.113578

Haifeng Li, Mo Hai

{"title":"PortfolioZero: A stock portfolio model based on deep reinforcement learning","authors":"Haifeng Li, Mo Hai","doi":"10.1016/j.asoc.2025.113578","DOIUrl":"10.1016/j.asoc.2025.113578","url":null,"abstract":"<div><div>Current studies of portfolio mainly use reinforcement learning methods to build models aimed at achieving high investment returns while minimizing risks from market uncertainties. Two main issues will be considered: First, the complexity of financial markets makes it challenging to capture asset price change patterns. Second, current research assumes stock prices accurately show all asset information, and historical prices alone can predict future trends. However, numerous external factors can influence future judgments. We introduce PortfolioZero, a novel model to address these problems. PortfolioZero utilizes three connected deep neural networks combined with a Monte Carlo Tree to discover patterns of financial assets. In the representation network, a Transformer-based model is used to embed financial price data to capture temporal dynamics and potential correlations, providing richer feature representations; the prediction network and Monte Carlo Tree Search are redesigned to handle the continuous action space. Furthermore, we use the StructBERT model to process financial text data, extracting market information into sentiment scores, which are used to reconstruct two reward functions to capture dynamic changes of the financial market. In experiments conducted on the China A-share market, we compared our model with traditional portfolio methods and cutting-edge deep reinforcement learning algorithms. PortfolioZero achieved an average annualized return rate of 21.21% across three portfolio types, outperforming SARL by 20.64% and DDPG by 41.97%, while sentiment-enhanced reward functions improved average annualized return rate by 35% compared to basic reward.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113578"},"PeriodicalIF":7.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711631","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}

引用次数: 0

A novel variable-precision granular-ball fuzzy rough set and its application in feature subset selection 一种新的变精度颗粒球模糊粗糙集及其在特征子集选择中的应用

IF 7.2 1区计算机科学

Applied Soft Computing Pub Date : 2025-07-23 DOI: 10.1016/j.asoc.2025.113597

Yongxi Chen , Zhehuang Huang , Anhui Tan , Jinjin Li

{"title":"A novel variable-precision granular-ball fuzzy rough set and its application in feature subset selection","authors":"Yongxi Chen , Zhehuang Huang , Anhui Tan , Jinjin Li","doi":"10.1016/j.asoc.2025.113597","DOIUrl":"10.1016/j.asoc.2025.113597","url":null,"abstract":"<div><div>Granular-ball computing is an efficient and interpretable theoretical method for multi-granularity data processing. Traditional fuzzy rough set models have some limitations in multi-granularity generation and dynamic characterization. On one hand, most of them lack effective multi-granularity generation methods, and artificially constructed approach may lead to information loss. On the other hand, when the attribute set changes, most fuzzy rough set models exhibit limitations in accurately characterize granularity information changes, and fails to capture correlational changes between attributes. Additionally, these models often lack noise resistance capabilities and flexibility in handling various fuzzy decision scenarios. To overcome these limitations, this paper combines granular-ball computing with fuzzy rough sets, and proposes a new variable-precision granular-ball fuzzy rough set model (VPGBFRS). First, granular-ball fuzzy similarity relations and granular-ball fuzzy neighborhood are used to characterize the relationship between samples. On this basis, a pair of variable-precision granular-ball approximate operators are presented. Second, we construct a variable-precision multi-granularity dependency function to obtain richer classification information, and enhance the model’s ability to capture intrinsic data structures. Finally, we design a forward attribute reduction algorithm based on the variable-precision significance in the sense of remain the classification ability unchanged. Numerical experiments conducted on 12 datasets demonstrate that, compared with four state-of-the-art attribute reduction algorithms, the proposed model exhibits superior performance, achieving significant improvements in both classification accuracy and the size of selected attribute set.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"183 ","pages":"Article 113597"},"PeriodicalIF":7.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711629","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}

引用次数: 0