Expert Systems with Applications最新文献_第3页

Adaptive dynamic prediction mechanism and heuristic algorithm based fast threshold selection for reversible data hiding 基于自适应动态预测机制和启发式算法的可逆数据隐藏快速阈值选择

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128251

Fengyun Shi, Yi Zhao, Wen Han, Junxiang Wang

{"title":"Adaptive dynamic prediction mechanism and heuristic algorithm based fast threshold selection for reversible data hiding","authors":"Fengyun Shi, Yi Zhao, Wen Han, Junxiang Wang","doi":"10.1016/j.eswa.2025.128251","DOIUrl":"10.1016/j.eswa.2025.128251","url":null,"abstract":"<div><div>The Prediction Error Expansion (PEE) framework has been extensively studied in the field of Reversible Data Hiding (RDH). Prediction-Error Value Ordering (PEVO), based on the PEE framework significantly optimizes the embedding performance by exploiting the correlation between prediction errors. Nevertheless, this scheme merely exploits the correlation between multiple maximum/minimum prediction errors within a block, and fails to adequately consider the case of large fluctuations in prediction errors. Therefore, to further exploit the redundancy in texture images, a novel median prediction scheme is proposed in this paper. It utilizes the median value to calculate multiple values on both sides, and this scheme is more accurate for correlation analysis of fluctuating prediction errors. To adapt to various texture images, an adaptive prediction mechanism is proposed, which combines the median prediction and PEVO approaches to generate more embeddable prediction error pairs. Unlike PEVO that employs fixed prediction values for all blocks, the proposed scheme dynamically selects the appropriate prediction scheme for each block based on inter-correlation between target value and its neighboring pixels. Moreover, instead of traversing all possible candidates for block noise level thresholds to determine optimal number of error pairs to be embedded, a heuristic-based algorithm is proposed that quickly determines noise level thresholds using an objective function based on the embeddable prediction error proportion. Finally, experimental results show that the proposed scheme outperforms state-of-the-art works. For example, when the capacity is 20,000 bits, the Boat image achieves a PSNR of 55.52 dB, showing a gain of 0.24 dB compared to the best results in the literature.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"288 ","pages":"Article 128251"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169519","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

Convolutional fuzzy modules stacked deep residual system with application to classification problems 卷积模糊模块堆叠深度残差系统在分类问题中的应用

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128282

Yunxia Liu , Xiao Lu , Haixia Wang , Jianqiang Yi , Chengdong Li

{"title":"Convolutional fuzzy modules stacked deep residual system with application to classification problems","authors":"Yunxia Liu , Xiao Lu , Haixia Wang , Jianqiang Yi , Chengdong Li","doi":"10.1016/j.eswa.2025.128282","DOIUrl":"10.1016/j.eswa.2025.128282","url":null,"abstract":"<div><div>Recent years have witnessed tremendous efforts devoted to investigating various data-driven methods, but how to build deep fuzzy models with good interpretability, high-precision, and well generalization ability remains a huge challenge, especially when facing complex, high-dimensional, and strong-nonlinear characteristics in the classification problems. Integrating both the advantages of convolutional neural networks and fuzzy inference method, this paper proposes a deep residual system by stacking the convolutional fuzzy modules (CFM-DRS), which achieves excellent performance with four mechanisms. Firstly, this study designs a new convolutional fuzzy module (CFM), which can comprehensively extract features from datasets with the convolutional operations, and then classify them through the corresponding sub-fuzzy-inference-modules (s-FIM). It is also the foundation of the other three mechanisms. Furthermore, each s-FIM employs the fuzzy C-means algorithm to identify the distribution patterns of features. It not only establishes the inference relationship between the features and output values in an interpretable manner, but also alleviates the problem of rule explosion. In addition, to reduce the impact of outliers and redundancy information on the overall performance, this study adopts the regularization optimization strategy to punish the parameters and prunes the s-FIMs based on their significant contributions. Besides, the utilization of the residual approximation mechanism in the deep framework is beneficial for learning new features and further improving the model’s accuracy. The proposed CFM-DRS is applied to several classification problems. Extensive experiments on different benchmark and real-world datasets demonstrate that the proposed CFM-DRS has a better classification performance compared to several state-of-the-art methods.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"288 ","pages":"Article 128282"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144169520","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

Correlation-assisted spatio-temporal reinforcement learning for stock revenue maximization 股票收益最大化的相关辅助时空强化学习

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128361

Jaehyun Chung , Minjoo Kim , Seokhyeon Min , Hyunseok Choi , Soohyun Park , Joongheon Kim

{"title":"Correlation-assisted spatio-temporal reinforcement learning for stock revenue maximization","authors":"Jaehyun Chung , Minjoo Kim , Seokhyeon Min , Hyunseok Choi , Soohyun Park , Joongheon Kim","doi":"10.1016/j.eswa.2025.128361","DOIUrl":"10.1016/j.eswa.2025.128361","url":null,"abstract":"<div><div>Investors struggle with the unpredictable, nonlinear nature of stock price volatility. Econometric models based on machine learning algorithms have improved prediction accuracy but remain limited in dynamic and highly correlated markets. This paper builds upon the proximal policy optimization (PPO) algorithm, the well-established deep reinforcement learning (DRL) method, and proposes an enhanced variant called correlation graph-based PPO (CGPPO), which incorporates spatio-temporal stock correlations for more realistic and robust predictions. The reward function, designed based on trading frequency and portfolio value, enhances experimental sophistication by reflecting practical investment objectives. The experiment is conducted in the simulated market environment using four major Korean stocks while explicitly considering the correlations among them. Experimental results show that the proposed CGPPO algorithm outperforms baseline methods, achieving <span><math><mrow><mn>64.60</mn><mspace></mspace><mo>%</mo></mrow></math></span> reward convergence value during training and <span><math><mrow><mn>69.04</mn><mspace></mspace><mo>%</mo></mrow></math></span> prediction value during inference.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128361"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144204544","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

Meta-Tuner: Meta-trained node-specific transformations for Graph Few-Shot Class-Incremental Learning 元调谐器：用于图少射类增量学习的元训练节点特定转换

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128332

Zhengnan Li , Jun Fang , Junbo Wang , Xilong Cheng , Yuting Tan , Yunxiao Qin

{"title":"Meta-Tuner: Meta-trained node-specific transformations for Graph Few-Shot Class-Incremental Learning","authors":"Zhengnan Li , Jun Fang , Junbo Wang , Xilong Cheng , Yuting Tan , Yunxiao Qin","doi":"10.1016/j.eswa.2025.128332","DOIUrl":"10.1016/j.eswa.2025.128332","url":null,"abstract":"<div><div>Recommender systems leverage Graph Neural Networks (GNNs) to model user-item interactions and learn high-quality representations, incorporating category priors for personalized recommendations. However, real-world graphs continuously evolve, introducing new categories with scarce labeled data, such as emerging music genres or newly introduced product types. As new items and user preferences emerge, models must effectively adapt to novel classes with only a few labeled examples. This motivates us to investigate the <strong>G</strong>raph <strong>F</strong>ew-<strong>S</strong>hot <strong>C</strong>lass-<strong>I</strong>ncremental <strong>L</strong>earning (GFSCIL) problem, where graph data continuously evolve with the emergence of new classes, each containing only a few labeled nodes. To address this, we introduce Meta-Tuner, a novel framework that meta-trains a GNN-based encoder and a hypernetwork on simulated GFSCIL tasks. The encoder extracts initial node embeddings, while the hypernetwork applies node-specific transformations to refine these embeddings. We also propose an Equalized Prototypical Distribution Loss (EPDL) to constrain prototypes on a regular n-simplex, encouraging the model to fully utilize the high-dimensional embedding space. Our experimental results across three datasets show that Meta-Tuner significantly outperforms current state-of-the-art methods, achieving clearer decision boundaries and up to 30 % accuracy gains in the GFSCIL problem. Moreover, by applying Meta-Tuner to existing few-shot node classification (FSNC) methods, we also achieve improved FSNC performance, highlighting Meta-Tuner’s effectiveness. <span><span>Code is available.</span><svg><path></path></svg></span></div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128332"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144204541","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

XGFu: Enhancing low-light visualization by feature and graph fusion of multiple artificial exposure images XGFu：通过多张人工曝光图像的特征和图形融合增强弱光可视化

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128308

Sihai Qiao , Tong Wang , Ming An , Rong Chen , Yushi Li

{"title":"XGFu: Enhancing low-light visualization by feature and graph fusion of multiple artificial exposure images","authors":"Sihai Qiao , Tong Wang , Ming An , Rong Chen , Yushi Li","doi":"10.1016/j.eswa.2025.128308","DOIUrl":"10.1016/j.eswa.2025.128308","url":null,"abstract":"<div><div>Images taken in low-light conditions often have a low dynamic range and include noise; however, existing multi-exposure image fusion methods are frequently affected by color and exposure levels, further complicating the saturation and dynamic range for high-quality images. In addressing these challenges, this paper introduces a two-dimensional graph convolutional multi-exposure image fusion framework (XGFu). It incorporates spatial and channel graph feature fusion for the feature fusion of artificially generated multi-exposure images. Specifically, the proposed enhancement framework consists of two networks: the Multi-Exposure Generation Network (MEG-Net) and the Graph-based Feature Fusion Network (GFF-Net). The MEG-Net combines nonlinear factor estimation modules with weighting schemes to generate a series of artificial exposure images, significantly expanding the hidden feature space of input low-light images. The GFF-Net supports both single-feature set depth fusion and multi-feature set breadth fusion of exposure-related feature sets, with both employing graph convolution to construct a joint reasoning chain of different feature graphs in spatial and channel dimensions. Qualitative and quantitative evaluations conducted on synthetic and real low-light images demonstrate that our model outperforms other state-of-the-art (SoTA) methods in robust low-light enhancement.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128308"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144204543","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

SemTG-Track: Multimodal fine-grained semantic-unit temporal guidance for multi-object tracking SemTG-Track：用于多目标跟踪的多模态细粒度语义单元时间制导

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128359

Kai Ren , Chuanping Hu , Hao Xi , Yongqiang Li , Jinhao Fan , Lihua Liu

{"title":"SemTG-Track: Multimodal fine-grained semantic-unit temporal guidance for multi-object tracking","authors":"Kai Ren , Chuanping Hu , Hao Xi , Yongqiang Li , Jinhao Fan , Lihua Liu","doi":"10.1016/j.eswa.2025.128359","DOIUrl":"10.1016/j.eswa.2025.128359","url":null,"abstract":"<div><div>In multi-object tracking (MOT) tasks, maintaining long-term identity consistency of targets in complex scenes remains a challenging research problem. Traditional prediction methods based on visual appearance features and motion trajectories struggle to dynamically and continuously preserve the unique representation of targets in complex environments. This limitation leads to tracking drift and identity confusion when targets undergo occlusion, blurring, or changes in scene dynamics and motion patterns, significantly degrading tracking performance. To address this issue, we propose a novel approach, SemTG-Track, which links the same target through cross-modal semantic information. By integrating a vision-language model with a hybrid LoRA expert system, our method enhances tracking performance through fine-grained modality alignment and dynamic semantic matching strategies. The SemTG-Track framework consists of three core modules: Semantic-unit Temporal Completeness Generation (STCG), Heterogeneous Semantic Representation Alignment (HSRA), and Temporal Sampling and Dynamic Matching (TSDM). Specifically, the STCG module leverages a vision-language model to generate a rich semantic knowledge graph for targets, the HSRA module enhances the generalization capability of semantic units through a dual-domain expert semantic fusion mechanism, and the TSDM module improves the efficiency and accuracy of multi-object tracking via dynamic sampling and context-aware matching mechanisms. Experimental results demonstrate that the proposed method outperforms baseline approaches, achieving improvements of 2.0 and 4.1 percentage points in MOTA and HOTA, respectively, on the MOT17 dataset. On the MOT20 dataset, our method also achieves gains of 0.4 and 2.2 percentage points in MOTA and HOTA, respectively, validating the effectiveness of our approach.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128359"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144185101","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

Hybrid iForest-DBSCAN for anomaly detection and wind power curve modelling 混合ifforest - dbscan用于异常检测和风力曲线建模

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128381

Zahid Mehmood, Zhenyu Wang

{"title":"Hybrid iForest-DBSCAN for anomaly detection and wind power curve modelling","authors":"Zahid Mehmood, Zhenyu Wang","doi":"10.1016/j.eswa.2025.128381","DOIUrl":"10.1016/j.eswa.2025.128381","url":null,"abstract":"<div><div>To achieve optimal performance and reduce the maintenance cost of wind turbines, anomaly detection and power curve modelling are crucial. Supervisory control and data acquisition (SCADA) data were collected from multiple wind farms with capacities of 4 MW and 1.5 MW, respectively. The SCADA data covered diverse operational scenarios of wind turbines with fluctuating wind speeds, rotor speeds, and blade pitches. This study introduces a novel strategy that combines the strengths of Isolation Forest (iForest) and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to identify and isolate anomalous data. The hybrid iForest-DBSCAN model processes enormous amounts of SCADA data to detect outliers and anomalies in wind turbines under various operating conditions. By utilizing normal data with minimum anomalies, normal behavioral power curves (NBPC) were modelled using a robust Locally Estimated Scattered Smoothing (LOESS) technique. Robust power curves allow us to compare the performances of wind turbines and ensure an optimized function with minimum maintenance. Different datasets validated the proposed method with a precision of 0.98 and higher accuracies of 99.8 % and 99.1 % in offshore wind farms 1 and 2 and onshore wind farms with 99.9 % precision and accuracy, respectively. Moreover, the computational resource requirements for this method in handling large datasets are minimal compared to traditional methods, such as neural-network-based methods. The iForest-DBSCAN model effectively identified anomalies from datasets obtained from the three wind farms, while successfully generating NBPC with a 95 % confidence interval. This study provides a data-driven framework based on real-time operating wind farms to optimize the operational efficiency, predictive maintenance, and enhance the lifespan of wind farms.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128381"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144189676","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

Explainable AI: XAI-guided context-aware data augmentation 可解释的AI: xai引导的上下文感知数据增强

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128364

Melkamu Abay Mersha , Mesay Gemeda Yigezu , Atnafu Lambebo Tonja , Hassan Shakil , Samer Iskander , Olga Kolesnikova , Jugal Kalita

{"title":"Explainable AI: XAI-guided context-aware data augmentation","authors":"Melkamu Abay Mersha , Mesay Gemeda Yigezu , Atnafu Lambebo Tonja , Hassan Shakil , Samer Iskander , Olga Kolesnikova , Jugal Kalita","doi":"10.1016/j.eswa.2025.128364","DOIUrl":"10.1016/j.eswa.2025.128364","url":null,"abstract":"<div><div>Explainable AI (XAI) has emerged as a powerful tool for improving the performance of AI models, going beyond providing model transparency and interpretability. The scarcity of labeled data remains a fundamental challenge in developing robust and generalizable AI models, particularly for low-resource languages. Conventional data augmentation techniques introduce noise, cause semantic drift, disrupt contextual coherence, lack control, and lead to overfitting. To address these challenges, we propose XAI-Guided Context-Aware Data Augmentation. This novel framework leverages XAI techniques to modify less critical features while selectively preserving most task-relevant features. Our approach integrates an iterative feedback loop, which refines augmented data over multiple augmentation cycles based on explainability-driven insights and the model performance gain. Our experimental results demonstrate that XAI-SR-BT and XAI-PR-BT improve the accuracy of models on hate speech and sentiment analysis tasks by 6.6 % and 8.1 %, respectively, compared to the baseline, using the Amharic dataset with the XLM-R model. XAI-SR-BT and XAI-PR-BT outperform existing augmentation techniques by 4.8 % and 5 %, respectively, on the same dataset and model. Overall, XAI-SR-BT and XAI-PR-BT consistently outperform both baseline and conventional augmentation techniques across all tasks and models. This study provides a more controlled, interpretable, and context-aware solution to data augmentation, addressing critical limitations of existing augmentation techniques and offering a new paradigm shift for leveraging XAI techniques to enhance AI model training.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 128364"},"PeriodicalIF":7.5,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144204283","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

Framework to automatically determine the quality of open data catalogs 框架自动确定开放数据目录的质量

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-29 DOI: 10.1016/j.eswa.2025.128379

Jorge Martinez-Gil

引用次数: 0

ClusterHopper: Cross-region order dispatching optimization for ride-hailing drivers ClusterHopper：网约车司机跨区域订单调度优化

IF 7.5 1区计算机科学

Expert Systems with Applications Pub Date : 2025-05-28 DOI: 10.1016/j.eswa.2025.127878

Yue Sun , Sasa Duan , Joseph Yen , Wen Xiong , Ming Jin , Yang Wang

{"title":"ClusterHopper: Cross-region order dispatching optimization for ride-hailing drivers","authors":"Yue Sun , Sasa Duan , Joseph Yen , Wen Xiong , Ming Jin , Yang Wang","doi":"10.1016/j.eswa.2025.127878","DOIUrl":"10.1016/j.eswa.2025.127878","url":null,"abstract":"<div><div>Current ride-hailing platforms operate in isolation, forcing drivers to choose between income stability and maximum earnings. This fragmented approach leads to inefficient resource allocation, with 30%–40% of driver time spent idle in low-demand areas while nearby regions experience order backlogs. We present ClusterHopper, a multi-region dispatching platform that coordinates orders across competing regions while preserving the platform autonomy for each region. By modeling each region as an independent matching queue and implementing a two-tier optimization framework based on the network flow principle with future revenue projection for order dispatch across different regions, our solution addresses four critical industry challenges in an incremental fashion for optimal global resource allocations: (1) System receipt revenue, (2) Quantity of completed orders, (3) Average waiting time, and (4) Average revenue. Compared to a single platform, the multi-platform approach (across four platforms) improved the four metrics by 47.05%, 31.24%, 61.36%, and 13.12%, respectively. Compared to ClusterHopper in its ride-hailing mode, its ride-sharing extension achieves improvements of 334.08%, 313.79%, 40.30%, and 3.55% across the four key metrics. In addition, the order cancellation rate was reduced by 68.35%. Real-world simulations using Didi’s operational data demonstrate consistent performance across varying demand patterns, proving the viability of cooperative competition in mobility markets.</div></div>","PeriodicalId":50461,"journal":{"name":"Expert Systems with Applications","volume":"289 ","pages":"Article 127878"},"PeriodicalIF":7.5,"publicationDate":"2025-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144189672","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