International Journal of Approximate Reasoning最新文献_第4页

On exact regions between measures of concordance and Chatterjee’s rank correlation for lower semilinear copulas 关于下半线性联结的一致性测度与查特吉等级相关之间的精确区域

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-11-02 DOI: 10.1016/j.ijar.2025.109588

Sebastian Fuchs, Carsten Limbach, Fabian Schürrer

{"title":"On exact regions between measures of concordance and Chatterjee’s rank correlation for lower semilinear copulas","authors":"Sebastian Fuchs, Carsten Limbach, Fabian Schürrer","doi":"10.1016/j.ijar.2025.109588","DOIUrl":"10.1016/j.ijar.2025.109588","url":null,"abstract":"<div><div>We explore how the classical concordance measures–Kendall’s <span><math><mi>τ</mi></math></span>, Spearman’s rank correlation <span><math><mi>ρ</mi></math></span>, and Spearman’s footrule <span><math><mi>ϕ</mi></math></span>–relate to Chatterjee’s rank correlation <span><math><mi>ξ</mi></math></span> when restricted to lower semilinear copulas. First, we provide a complete characterization of the attainable <span><math><mrow><mi>τ</mi><mspace></mspace><mo>−</mo><mspace></mspace><mi>ρ</mi></mrow></math></span> region for this class, thus resolving the conjecture of Maislinger and Trutschnig [1]. Building on this result, we then derive the exact <span><math><mrow><mi>τ</mi><mspace></mspace><mo>−</mo><mspace></mspace><mi>ϕ</mi></mrow></math></span> and <span><math><mrow><mi>ϕ</mi><mspace></mspace><mo>−</mo><mspace></mspace><mi>ρ</mi></mrow></math></span> regions, obtain a closed-form relationship between <span><math><mi>ξ</mi></math></span> and <span><math><mi>τ</mi></math></span>, and establish the exact <span><math><mrow><mi>τ</mi><mspace></mspace><mo>−</mo><mspace></mspace><mi>ξ</mi></mrow></math></span> region. In particular, we prove that <span><math><mi>ξ</mi></math></span> never exceeds <span><math><mi>τ</mi></math></span>, <span><math><mi>ρ</mi></math></span>, or <span><math><mi>ϕ</mi></math></span>. Our results clarify the relationship between undirected and directed dependence measures and reveal novel insights into the dependence structures that result from lower semilinear copulas.</div></div>","PeriodicalId":13842,"journal":{"name":"International Journal of Approximate Reasoning","volume":"189 ","pages":"Article 109588"},"PeriodicalIF":3.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145464607","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Matrix-based efficient methods to update three-way regions in neighborhood systems under varying attributes 基于矩阵的邻域系统三向区域更新方法

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-11-10 DOI: 10.1016/j.ijar.2025.109598

Yingqi Qi , Chengxiang Hu , Xiaoling Huang

{"title":"Matrix-based efficient methods to update three-way regions in neighborhood systems under varying attributes","authors":"Yingqi Qi , Chengxiang Hu , Xiaoling Huang","doi":"10.1016/j.ijar.2025.109598","DOIUrl":"10.1016/j.ijar.2025.109598","url":null,"abstract":"<div><div>Traditional set-based methods for computing three-way regions in neighborhood systems primarily rely on the inclusion relationships between target concepts and neighborhood classes to process continuous numerical data. However, these methods exhibit significant limitations when applied to time-varying neighborhood information systems, as they inherently lack the capability to accommodate dynamically evolving data, effectively. To overcome this challenge, our research presents novel matrix-based incremental methods that leverage previously computed results to enable more efficient updating and maintenance of three-way regions in neighborhood rough sets. Through comprehensive integration and analysis of neighborhood information systems with a focus on varying attributes, we develop matrix-based incremental mechanisms. Building on these mechanisms, we propose two incremental algorithms to effectively handle dynamic numerical data. Experimental results demonstrate the effectiveness and superior efficiency of the proposed methods compared to existing approaches. Specifically, the proposed algorithms exhibit lower computational time and higher speed-up ratio, highlighting their efficiency for updating neighborhood three-way regions.</div></div>","PeriodicalId":13842,"journal":{"name":"International Journal of Approximate Reasoning","volume":"189 ","pages":"Article 109598"},"PeriodicalIF":3.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145621173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Disentangled representations for continuous treatment effect estimation 连续处理效果估计的解纠缠表示

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-11-16 DOI: 10.1016/j.ijar.2025.109601

Ruijing Cui, Jianbin Sun, Bingyu He, Kewei Yang, Bingfeng Ge

引用次数: 0

Feature enhancement-based network for few-shot image classification 基于特征增强的少拍图像分类网络

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-11-19 DOI: 10.1016/j.ijar.2025.109605

Junbo Zhao , Lixin Han , Hong Yan

{"title":"Feature enhancement-based network for few-shot image classification","authors":"Junbo Zhao , Lixin Han , Hong Yan","doi":"10.1016/j.ijar.2025.109605","DOIUrl":"10.1016/j.ijar.2025.109605","url":null,"abstract":"<div><div>Few-shot image classification tackles recognizing new classes with limited training data. Despite the progress made by researchers in recent years, there are still challenges in dealing with data scarcity and improving model robustness. This paper presents FEICNet, a framework combining feature enhancement and adaptive weight control. Key innovations include: (1) MiniUNet (MUNet) — a compact network that enhances spatial and channel features through layered processing, and (2)Distribution of Feature Weights (DFW) – an adaptive weighting system that boosts critical features while filtering irrelevant patterns. Compatible with existing models without structural changes, the framework achieves strong performance across five benchmarks (MiniImageNet, tieredImageNet, StandfordDogs, CUB, StanfordCars). On StanfordCars, FEICNet outperforms Conv-64F methods by 13.4 % (1-shot) and 13.5 % (5-shot), surpassing ResNet-12 models by 9.7 % in 5-shot tests. Notably, when integrated as an embedding module, FEICNet elevates ProtoNets and Relation Networks by 19.6 %–28.8 % on StanfordDogs, demonstrating its effective plug-and-play capability. The framework exhibits consistent convergence across episodic training tasks, further evidencing its robustness. These advancements establish FEICNet as a significant contribution to few-shot image classification, particularly in resource-constrained environments and coarse- and fine-grained recognition scenarios. Overall, FEICNet provides a unified framework that bridges feature enhancement with adaptive weighting, offering new insights into few-shot image representation learning.</div></div>","PeriodicalId":13842,"journal":{"name":"International Journal of Approximate Reasoning","volume":"189 ","pages":"Article 109605"},"PeriodicalIF":3.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145621174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

New paraconsistent modal logics based on rough modus ponens rules and their interrelations 基于粗糙模态规则及其相互关系的新准一致模态逻辑

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-11-01 DOI: 10.1016/j.ijar.2025.109587

Bidhan Saha , Mohua Banerjee , Soma Dutta

引用次数: 0

Modeling the selection of representative aggregation functions for optimizing the representation of group behavior preferences within the preference disaggregation framework 在偏好分解框架下，为优化群体行为偏好的表示，对代表性聚合函数的选择进行建模

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-02-01 Epub Date: 2025-10-26 DOI: 10.1016/j.ijar.2025.109586

Zaiwu Gong , Jiaxin Zhou , Xinxin Luo , Kun Zhou , Weiwei Guo , Guo Wei

{"title":"Modeling the selection of representative aggregation functions for optimizing the representation of group behavior preferences within the preference disaggregation framework","authors":"Zaiwu Gong , Jiaxin Zhou , Xinxin Luo , Kun Zhou , Weiwei Guo , Guo Wei","doi":"10.1016/j.ijar.2025.109586","DOIUrl":"10.1016/j.ijar.2025.109586","url":null,"abstract":"<div><div>In Multiple Criteria Decision Aid (MCDA), preference disaggregation is used to derive aggregation functions that replicate the preference information provided by decision makers (DMs). A key research challenge is identifying the most representative aggregation function that aligns with the DMs’ behavioral preferences from all feasible functions. In the context of group decisions and negotiations, preference disaggregation faces two main barriers: Selecting representative aggregation functions and reducing the cognitive effort required from DMs. In this paper, we extend preference disaggregation to group decision-making by enabling DMs to increase the diversity of their decision matrices, and establish several representative aggregation function selection models that account for group behavioral preferences. Given the cognitive limitations of DMs, their preferences are often expressed as confidence levels, which inherently involve uncertainty. Therefore, we expand the scope of preference information required for preference disaggregation from deterministic preferences to uncertain preferences. By applying probability theory–specifically, chance constraints–we transform DMs’ uncertain preferences into constraints for selecting representative aggregation functions that align with their behavioral preferences. Notably, when the confidence level approaches 100 %, the uncertain preferences converge to their deterministic counterparts, demonstrating the extensiveness and inclusiveness of uncertain preference information. Finally, we illustrate the practical feasibility of our approach through a case study on the evaluation of medical pension institutions.</div></div>","PeriodicalId":13842,"journal":{"name":"International Journal of Approximate Reasoning","volume":"189 ","pages":"Article 109586"},"PeriodicalIF":3.0,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145420387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Special issue on the 11th International Conference on Soft Methods in Probability and Statistics (SMPS 2024) 第十一届国际概率与统计软方法会议（SMPS 2024）特刊

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-01-01 Epub Date: 2025-10-09 DOI: 10.1016/j.ijar.2025.109584

Sebastian Fuchs

引用次数: 0

Zero patterns in multi-way binary contingency tables with uniform margins 等边距多路二元列联表中的零模式

IF 3 3区计算机科学

International Journal of Approximate Reasoning Pub Date : 2026-01-01 Epub Date: 2025-10-02 DOI: 10.1016/j.ijar.2025.109582

Roberto Fontana , Elisa Perrone , Fabio Rapallo

引用次数: 0

Granular structures of covering-based rough set approximations through the Mapper algorithm 通过Mapper算法实现基于覆盖的粗糙集近似的颗粒结构