Information Sciences最新文献_第5页

Predicting cranial lesion evolution with temporal attention 用时间注意力预测颅脑损伤演变

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-15 DOI: 10.1016/j.ins.2025.122300

Riadh Bouslimi , Mariem Medini , Wahiba Ben Abdsalem Karaa , Hana Hedhli , Safia Othmani

{"title":"Predicting cranial lesion evolution with temporal attention","authors":"Riadh Bouslimi , Mariem Medini , Wahiba Ben Abdsalem Karaa , Hana Hedhli , Safia Othmani","doi":"10.1016/j.ins.2025.122300","DOIUrl":"10.1016/j.ins.2025.122300","url":null,"abstract":"<div><div>Anticipating the trajectory of cranial lesion progression with high precision represents a formidable challenge in neurosurgery, where the capacity for timely and accurate prognostication is indispensable for optimizing patient care strategies. This study introduces an advanced framework underpinned by a Seq2Seq architecture, augmented with a sophisticated temporal attention mechanism specifically designed for the analysis of longitudinal clinical datasets. The proposed architecture seamlessly integrates LSTM networks with Transformer-based components within a dynamic and adaptable decoder, thereby enabling the model to capture both transient fluctuations and enduring trends across complex patient data. To ensure alignment with practical clinical needs, a tailored loss function was developed, emphasizing critical lesion parameters, including volumetric dimensions, morphological complexity, and signal intensity. Rigorous validation was conducted using publicly available datasets, such as MIMIC-IV and eICU, where the model consistently outperformed traditional methodologies, including RNNs and standard Transformer frameworks. Moreover, the inclusion of the temporal attention mechanism enhances interpretability by identifying critical temporal windows associated with key clinical events. These results underscore the transformative potential of artificial intelligence to generate dynamic, context-aware insights, facilitating personalized treatment planning despite the inherent complexity and variability of clinical datasets.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122300"},"PeriodicalIF":8.1,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144083745","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 granular-ball generation method based on local density for classification 基于局部密度的颗粒球生成方法进行分类

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-15 DOI: 10.1016/j.ins.2025.122295

Fan Liu , Qinghua Zhang , Shuyin Xia , Qin Xie , Wei Liao , Siyang Zhang

{"title":"A granular-ball generation method based on local density for classification","authors":"Fan Liu , Qinghua Zhang , Shuyin Xia , Qin Xie , Wei Liao , Siyang Zhang","doi":"10.1016/j.ins.2025.122295","DOIUrl":"10.1016/j.ins.2025.122295","url":null,"abstract":"<div><div>As a new branch of granular computing, granular-ball computing (GBC) has become increasingly popular due to its high efficiency, robustness, and scalability. However, in classification tasks, the existing mainstream methods for generating granular-ball (GB) have two common issues: GBs generated by the existing methods are not accurate enough to describe the distribution of the original data and there are de-overlap operations in the existing GB generation process, which increase the workload of the GB generation process. Therefore, to solve the above two issues, a GB generation method based on local density (LDGBG) is proposed in this paper. First, centers of GBs are selected based on local density to ensure the generated GBs are more consistent with the original data distribution. Second, the method for calculating radii avoids overlaps and incorporates the idea of compact within class and decentralized between classes, which will improve classification performance. Furthermore, a sparsity index is introduced to assess the sparsity of datasets, thereby enabling more effective utilization of original samples in sparse datasets. Finally, comparative experiments are conducted on 27 benchmark datasets and the experimental results show that LDGBG is superior to the existing mainstream models in effectiveness and robustness.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122295"},"PeriodicalIF":8.1,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144083744","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

RsDiff: Rational score based knowledge graph diffusion for recommendation RsDiff：基于理性分数的知识图谱扩散推荐

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122292

Mengmeng Cui , Siyu Wu , Hao Chen , Xiangnan Zhang

引用次数: 0

X-clustering beyond contextual representations 超越上下文表示的x聚类

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122291

Tianyi Huang , Zhengjun Zhang , Xin Yuan , Stan Z. Li , Neal Naixue Xiong , Shenghui Cheng

{"title":"X-clustering beyond contextual representations","authors":"Tianyi Huang , Zhengjun Zhang , Xin Yuan , Stan Z. Li , Neal Naixue Xiong , Shenghui Cheng","doi":"10.1016/j.ins.2025.122291","DOIUrl":"10.1016/j.ins.2025.122291","url":null,"abstract":"<div><div>Despite the success of clustering and eXplainable AI (XAI), most clustering methods lack explainability. Traditional clustering generates pseudo labels but ignores probability distributions, cluster similarities, and local manifold structures, making the process and results difficult to interpret. Deep clustering improves this by jointly learning probability distributions and representations, but the learned representations often lack contextual meaning and misalign with data topology.</div><div>To address this, we propose <em>X-Clustering beyond Contextual Representations (XCR)</em>. It learns contextual representations and maps them for the visual explanation of the clustering probability distributions, cluster similarities, and topological structure of each cluster. Samples with higher cluster probabilities are positioned closer to corresponding anchors. Using an adjacency graph, connected samples stay near the same anchor with similar probabilities, while unconnected samples remain distant. Anchors are adaptively updated to index clustering distributions accurately. Finally, samples and anchors are fused into a unified space for intuitive visual explanations. Experiments demonstrate that compared with existing clustering methods, XCR not only provides better explainable visual clustering results but also achieves better average ACC and NMI on five testing datasets. Furthermore, the average sample-cluster tagging accuracy of anchors on five testing datasets is 97.3% with the visualization of UMAP.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122291"},"PeriodicalIF":8.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068509","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

Evaluating supply chain finance risks in a cross-border e-commerce context: An improved TOPSIS approach with loss penalty 跨境电子商务环境下供应链金融风险评估：一种考虑损失惩罚的改进TOPSIS方法

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122301

Jinzhao Shi , Maolin Sun , Xiao Yang , Kewen Jing , Kin Keung Lai

{"title":"Evaluating supply chain finance risks in a cross-border e-commerce context: An improved TOPSIS approach with loss penalty","authors":"Jinzhao Shi , Maolin Sun , Xiao Yang , Kewen Jing , Kin Keung Lai","doi":"10.1016/j.ins.2025.122301","DOIUrl":"10.1016/j.ins.2025.122301","url":null,"abstract":"<div><div>The rapid development of cross-border e-commerce (CBEC) has created urgent demands for efficient capital coordination among cross-border supply chain members. Here, CBEC-supply chain finance (CBEC-SCF) is considered an effective solution. Based on the classification of traditional supply chain finance and CBEC characteristics, this study systematically proposes three categories of operational modes for CBEC-SCF: CBEC-based warehouse receipt financing, order financing, and factoring. Then, CBEC-SCF risks are analyzed from various perspectives, including credit, market, operational, and legal risks. Evaluating the overall risk levels of different CBEC-SCF modes is a typical multi-criteria decision-making (MCDM) problem, where various sub-risks serve as criteria. Given that decision makers (e.g., banks) are usually loss-averse, this study proposes an improved Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method with a loss penalty to rank the overall risk levels of different CBEC-SCF modes. The boundary conditions under which the improved method changes the ranking of alternatives are theoretically proven. To better fit a wider range of application scenarios, the improved TOPSIS is further extended to cases involving multi-level criteria, “experts” as the criteria, and fuzzy decision-making. Finally, case studies are conducted to verify the proposed method’s applicability and significance. The results show that the ranking of the CBEC-SCF modes may change depending on the bank’s degree of loss aversion.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122301"},"PeriodicalIF":8.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144088891","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 new lightweight convolutional neural network model for detecting drivable road regions 一种新的用于可行驶道路区域检测的轻量级卷积神经网络模型

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122305

Gürkan Doğan , Hakan Uyanık , Burhan Ergen

{"title":"A new lightweight convolutional neural network model for detecting drivable road regions","authors":"Gürkan Doğan , Hakan Uyanık , Burhan Ergen","doi":"10.1016/j.ins.2025.122305","DOIUrl":"10.1016/j.ins.2025.122305","url":null,"abstract":"<div><div>Nowadays, due to the rapid increase in the number of autonomous vehicles on the market, the safe navigation of these vehicles in drivable road areas has become extremely important. One of the most crucial factors in ensuring safe navigation is addressing the detection of drivable road areas as a task of semantic segmentation. Considering that autonomous vehicles are modular, the algorithm to perform this task must have the optimum trade-off in terms of lightweight, computational complexity, and segmentation accuracy. In this study, RoNet, a new model based on convolutional neural networks that provides an optimum trade-off for the detection of drivable road regions, was designed and proposed. The standard convolution types for the encoder and decoder bottleneck module of the RoNet model, as well as the spatial edge attention mechanism, have been optimized by developing asymmetric convolution types using asymmetric atrous convolution, asymmetric convolution types using Prewitt and Sobel kernels. Spatial edge attention mechanism is designed to reduce the loss of detailed information in small-resolution feature maps. In experimental tests performed with CamVid and FUVid datasets, RoNet achieved a better trade-off in terms of segmentation accuracy, number of parameters, and computational complexity compared to other state-of-the-art methods.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122305"},"PeriodicalIF":8.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144099894","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

GAL: A global aspect local extraction mechanism for aspect-based sentiment classification GAL：基于方面的情感分类的全局方面局部提取机制

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122299

Zhiqiang Zhang , Xiaoming Li , Hongpeng Bai , Meilian Zheng , Kun Huang

{"title":"GAL: A global aspect local extraction mechanism for aspect-based sentiment classification","authors":"Zhiqiang Zhang , Xiaoming Li , Hongpeng Bai , Meilian Zheng , Kun Huang","doi":"10.1016/j.ins.2025.122299","DOIUrl":"10.1016/j.ins.2025.122299","url":null,"abstract":"<div><div>Aspect-based sentiment classification (ABSC) is a subtask of aspect-based sentiment analysis (ABSA) that aims to predict the sentiment polarity of different aspects within a sentence. Existing research focuses primarily on combining two semantic features: the context of the sentence with the features of the aspect or the context of the sentence with local context features. However, these two-by-two combinations of extracting global or local contextual and aspectual features suffer from the limitation of insufficient semantic feature information and do not take full advantage of the feature information of the other in each piece of data. To address this issue, this paper presents a global aspect local (GAL) extraction mechanism designed to integrate global, aspect, and local context information to increase the accuracy of sentiment polarity prediction. This study designs and implements the GAL mechanism, which incorporates a multihead self-attention mechanism, local context focus, and a multilayer attention mechanism, utilizing pretrained models for text encoding. Experimental results on the SemEval-2014 Task 4 Subtask 2 dataset and the ACL Twitter social dataset demonstrate that our GAL mechanism improves accuracy by nearly 2% and the F1 score by 3% in fine-grained sentiment classification tasks compared with recent methods, verifying its effectiveness and superiority.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122299"},"PeriodicalIF":8.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068508","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

Optimizing decentralized peer-to-peer energy trading with energy-backed tokens through integration of renewable energy, battery storage, and electric vehicles 通过整合可再生能源、电池存储和电动汽车，优化能源支持代币的去中心化点对点能源交易

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-14 DOI: 10.1016/j.ins.2025.122294

Li Chang , Bingxiang Wu , Yiming He , Chao Zhou

{"title":"Optimizing decentralized peer-to-peer energy trading with energy-backed tokens through integration of renewable energy, battery storage, and electric vehicles","authors":"Li Chang , Bingxiang Wu , Yiming He , Chao Zhou","doi":"10.1016/j.ins.2025.122294","DOIUrl":"10.1016/j.ins.2025.122294","url":null,"abstract":"<div><div>Energy-backed token-based P2P trading reduces grid dependency, lowers costs, and boosts self-sufficiency. Integrating renewables, battery storage, and EVs further enhances system flexibility, supporting a sustainable energy transition. This study proposed a decentralized peer-to-peer energy-backed token market using smart contracts and a delay mechanism to limit inflation and future consumer collusion, which includes smart meters, electric vehicles, and electric vehicle charging stations. An optimized scheduling algorithm based on forecasted generation and demand was developed to increase individual and collective welfare. Smart parking spaces enabled electric vehicle participation in the market, while an ADMM algorithm facilitated decentralized market settlement based on the charging dynamics of electric vehicles. This work investigates one decentralized energy trading model with 36 participants (12 prosumers, 19 consumers, 5 EVs) using energy-backed tokens. Case 1 achieves 15.9% self-sufficiency; Case 2 adds EVs, boosting flexibility. Model meets 99% of DSO requests, cuts costs, and achieves 0.0017 validation and 0.0024 test performance in 38 epochs over 6 s. The model accurately forecasted photovoltaic power fluctuations, especially during peak times. Integrating token-based energy trading with adaptive battery management and EVs reduced energy costs and boosted user independence. Implementing a demurrage mechanism stabilized the token market and improved social welfare.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122294"},"PeriodicalIF":8.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144115017","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

Modeling linguistic intuitionistic fuzzy preference into the consensus and dissent framework of graph model for conflict resolution and its application 将语言直觉模糊偏好建模为冲突解决的共识与异议图模型框架及其应用

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-13 DOI: 10.1016/j.ins.2025.122288

Guolin Tang , Tangzhu Zhang , Yingting Lv , Peide Liu

{"title":"Modeling linguistic intuitionistic fuzzy preference into the consensus and dissent framework of graph model for conflict resolution and its application","authors":"Guolin Tang , Tangzhu Zhang , Yingting Lv , Peide Liu","doi":"10.1016/j.ins.2025.122288","DOIUrl":"10.1016/j.ins.2025.122288","url":null,"abstract":"<div><div>The intensification of water scarcity and pollution has elevated the strategic significance of cross-border water resources. Their transnational nature complicates ownership and use rights, leading to conflicts. The graph model for conflict resolution (GMCR) has proven effective in addressing strategic disputes. However, decision-makers' preferences, influenced by a complex interplay of cultural, economic, and political factors, are too intricate to quantify precisely in cross-border water resource conflicts. Besides, despite the importance of consensus being widely researched in the group decision-making field, its study within GMCR remains limited. To address these challenges, we develop a new consensus and dissent framework of GMCR with linguistic intuitionistic fuzzy preference relations (LIFPRs). As a qualitative tool, LIFPRs composed of linguistic intuitionistic fuzzy numbers (LIFNs) can capture DMs' certainty, uncertainty, and hesitation. Specifically, we first propose a score function of LIFNs and some necessary definitions to model LIFPRs into GMCR. Then, we study the consensus and dissent framework of GMCR with LIFPRs in logical form. Additionally, we provide the corresponding matrix representation for future decision support system development. This study is applied to the Mekong River conflict, suggesting that the Mekong River Commission and China should engage in a comprehensive cooperative negotiation process.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122288"},"PeriodicalIF":8.1,"publicationDate":"2025-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143949049","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

Attaining the stabilizing solution of model unavailable modified algebraic Riccati equation using Q-learning algorithm 利用q -学习算法获得模型不可用修正代数Riccati方程的稳定解

IF 8.1 1区计算机科学

Information Sciences Pub Date : 2025-05-13 DOI: 10.1016/j.ins.2025.122265

Jie Gao , Tao Feng , Fei Yan

{"title":"Attaining the stabilizing solution of model unavailable modified algebraic Riccati equation using Q-learning algorithm","authors":"Jie Gao , Tao Feng , Fei Yan","doi":"10.1016/j.ins.2025.122265","DOIUrl":"10.1016/j.ins.2025.122265","url":null,"abstract":"<div><div>This paper attempts first to solve the discrete-time modified algebraic Riccati equation (MARE) when the system model is completely unavailable. To achieve this, a new iterative algorithm is proposed to solve the MARE based on its equivalent discrete-time algebraic Riccati equation (DARE) which has an identical solution. The algorithm iterates between two steps: 1) the solution improvement of the equivalent DARE; 2) the input weighting improvement. It is proved theoretically that the algorithm strictly converges to the stabilizing solution of the MARE. As an outstanding advantage over the existing methods, the initiation of the algorithm becomes fairly simple which can be initialized by an arbitrary positive input weighting for the single-input scenario or a pre-given input weighting matrix of a sufficiently large magnitude for the multi-input scenario. Furthermore, the proposed algorithm provides an iterative structure which is particularly suitable for developing the Q-learning (QL) algorithm where only the input/output data of the related linear quadratic regulator (LQR) problem is required, therefore the MARE can be solved successfully without using the system model. Finally, numerical simulation examples are given to verify the effectiveness of the theoretical results and the developed algorithm.</div></div>","PeriodicalId":51063,"journal":{"name":"Information Sciences","volume":"717 ","pages":"Article 122265"},"PeriodicalIF":8.1,"publicationDate":"2025-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068506","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