检索结果-内蒙古大学图书馆

Enhancing Imbalance learning: A Novel Slack-Factor Fuzzy SVM Approach

IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE 2025年

作者： Tanveer, M. Tiwari, Anushka Akhtar, Mushir Lin, Chin-Teng Indian Inst Technol Indore Dept Math Simrol 453552 India Univ Technol Sydney Human Centr AI Ctr Sch Comp Sci Sydney NSW 2007 Australia

In real-world applications, class-imbalanced datasets pose significant challenges for machine learning algorithms, such as support vector machines (SVMs), particularly in effectively managing imbalance, noise, and outliers. Fuzzy support vector machines (FSVMs) address class imbalance by assigning varying fuzzy memberships to samples;however, their sensitivity to imbalanced datasets can lead to inaccurate assessments. The recently developed slack-factor-based FSVM (SFFSVM) improves traditional FSVMs by using slack factors to adjust fuzzy memberships based on misclassification likelihood, thereby rectifying misclassifications induced by the hyperplane obtained via different error cost (DEC). Building on SFFSVM, we propose an improved slack-factor-based FSVM (ISFFSVM) that introduces a novel location parameter. This novel parameter significantly advances the model by constraining the DEC hyperplane's extension, thereby mitigating the risk of misclassifying minority class samples. It ensures that majority class samples with slack factor scores approaching the location threshold are assigned lower fuzzy memberships, which enhances the model's discrimination capability. Extensive experimentation on a diverse array of real-world KEEL datasets demonstrates that the proposed ISFFSVM consistently achieves higher F1-scores, Matthews correlation coefficients (MCC), and area under the precision-recall curve (AUC-PR) compared to baseline classifiers. Consequently, the introduction of the location parameter, coupled with the slack-factor-based fuzzy membership, enables ISFFSVM to outperform traditional approaches, particularly in scenarios characterized by severe class disparity.

关键词： Frequency modulation Support vector machines Noise Vectors Costs Mathematical models Fasteners Computational modeling Computational intelligence machine learning algorithms Fuzzy membership (FM) support vector machine (SVM) different error cost (DEC) slack-factor-based fuzzy support vector machine (SFFSVM)

来源：评论

学校读者我要写书评

暂无评论

machine learning-inspired study of dynamical parameters of single vapor bubble under nucleate flow boiling regime

引用

APPLIED THERMAL ENGINEERING 2025年 259卷

作者： Moiz, Mohd Prajapat, Rakhee Srivastava, Arpan Srivastava, Atul Indian Inst Technol Dept Mech Engn Mumbai 400076 India Indian Inst Informat Technol Una Dept Elect & Commun Engn Saloh 177209 India Nirma Univ Inst Technol Dept Elect & Commun Ahmadabad 382481 India

Heat transfer performance of nucleate boiling is interlinked with the departure dynamics of vapor bubbles, which also serves as the basis for heat transfer partitioning schemes and bubble growth models. Given the fast transients and multiple bubble ebullition cycles, boiling experiments lead to voluminous data that are to be analysed for determining the bubble dynamic parameters, and subsequently, the boiling heat transfer rates. Conventional approach of manual handling of such temporal and spatially-resolved experimental data is inherently timeconsuming and error-prone. In the backdrop of these limitations, importance of machine learning algorithms gets highlighted in making the entire process automated and accurate as they minimize the need of any human intervention. The present work explores the potential of machine learning techniques towards quantifying the bubble departure characteristics and heat transfer partitioning during nucleate flow boiling. Water-based experiments have been conducted in a vertical channel under varying subcooling levels (Delta Tsub = 2, 5, and 8 K) and flow rates (Re = 2400, and 3600). Mask R-CNN machine learning (ML) model is employed to evaluate the temporal variations of dynamical parameters of vapor bubbles and departure characteristics. The bubble departure characteristics as well as the corresponding evaporative and convective heat transfer rates, as retrieved through ML-generated masks, showed a strong dependence on the level of bulk fluid subcooling and flow rates. Heat transfer partitioning analysis revealed a competing interplay between the evaporative and condensation components of overall boiling heat transfer rates. These findings demonstrate the potential of ML techniques towards optimizing the thermal performance of boiling phenomena that have significant implications in a range of critical engineering applications.

关键词： machine learning algorithms Mask R -CNN Nucleate flow boiling Bubble dynamic parameters Bulk subcooling levels

来源：评论

学校读者我要写书评

暂无评论

SASE-TCN: A Multitask learning Network for MDS-UPDRS III Task Classification and Parkinson's Disease Recognition

引用

IEEE SENSORS JOURNAL 2025年第8期25卷 13554-13565页

作者： Zhang, Yiyuan Meng, Long Chen, Chen Chen, Wei Fudan Univ Ctr Intelligent Med Elect Sch Informat Sci & Technol Shanghai 200433 Peoples R China Penn State Univ Dept Mech Engn University Pk PA 16802 USA Fudan Univ Shanghai Pudong Hosp Ctr Med Res & Innovat Pudong Med Ctr Shanghai 201399 Peoples R China Univ Sydney Sch Biomed Engn Sydney NSW 2050 Australia

As wearable technology and machine learning (ML) algorithms have developed by leaps and bounds, many studies have focused on automatically monitoring Parkinson's disease (PD). The procedure includes two steps: first, classifying activities, usually the Movement Disorder Society-sponsored revision of the unified PD rating scale (MDS-UPDRS) part III tasks;second, evaluating the performance of these activities and recognizing PD. To improve the efficiency and precision of the procedure, this study proposes a self-attention squeeze-and-excitation temporal convolutional network (SASE-TCN). Based on the multistage temporal convolutional network (MS-TCN) and attention mechanisms, SASE-TCN can effectively extract the distant sequential and channelwise features. In addition, based on acceleration signals, SASE-TCN can classify MDS-UPDRS III tasks and recognize patients with PD simultaneously with the multitask learning mechanism. The model was tested with two public datasets, PD-BioStampRC21 and PD-motion. The SASE-TCN model obtained an average F1 score of 0.7802-0.9160 for activity classification and an average accuracy of 0.7366-0.8578 for PD recognition. The results demonstrated the feasibility of the SASE-TCN in classifying PD-related activities and recognizing PD with accelerometers. This study will support the diagnosis and treatment of PD.

关键词： Convolutional neural networks Feature extraction Multitasking Diseases Data mining Motors machine learning algorithms Electronic mail Accuracy Accelerometers Attention mechanism multitask learning Parkinson's disease (PD) temporal convolutional network (TCN)

来源：评论

学校读者我要写书评

暂无评论

Stochastic analysis and machine learning techniques for maintenance of heritage buildings with multi-source data

引用

JOURNAL OF BUILDING ENGINEERING 2025年 103卷

作者： Pang, Bo Wang, Feiliang Zhang, Anshan Zhang, Kai Yang, Jian Shanghai Jiao Tong Univ State Key Lab Ocean Engn Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ Sch Ocean & Civil Engn Shanghai Key Lab Digital Maintenance Bldg & Infras Shanghai 200240 Peoples R China Ningbo Univ Fac Mech Engn & Mech Key Lab Impact & Safety Engn Minist Educ Ningbo 315211 Peoples R China Univ Birmingham Sch Civil Engn Birmingham B15 2TT England

The maintenance of buildings, enabled by a fusion of diverse techniques, becomes integral to heritage conservation efforts. According to the principle of minimal intervention, applied methodology employs constrained survey and monitoring, resulting in an information deficit that may challenge comprehensive structural assessments. To address the uncertainty introduced in assessing structural risk assessment and performance prediction, stochastic analysis based on reasonable assumptions is required. machine learning algorithms offer approaches to obtain valuable insights from limited datasets. This study presents a maintenance case where a monitoring system was deployed at a heritage site, underscoring the adaptability of these methods. The probability density evolution method (PDEM) and a credibility index were employed to quantify the possible failure and evolution of the City Wall stability, providing prior knowledge based on experience. Potential failure modes and unstable trends consistent with the observed results were revealed. The integration of machine learning techniques offers robust predictive models that address data scarcity and identify environmental influences. Vision and sound-based techniques also demonstrated significant promise in risk identification and forming multi-source archive for heritage maintenance.

关键词： Heritage buildings Structural health monitoring Uncertainty Prediction models machine learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Memory Priority Scheduling Algorithm for Cloud Data Center Based on machine learning Dynamic Clustering Algorithm

引用

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS 2025年第4期21卷 3485-3492页

作者： Liang, Bin Wu, Di Shiyou Univ Sch Comp Sci Xian 710065 Peoples R China Shaanxi Univ Sci & Technol Haojing Coll Xian 712046 Peoples R China

As the cloud data center (CDC) landscape continues to broaden, CDC resource utilization as the benchmark for assessing scheduling methodologies. Concurrently enhancing both CPU and memory utilization stands as a paramount priority. However, prevalent algorithms tend to solely prioritize CPU utilization while neglecting memory efficiency, ultimately escalating energy expenditure. This article initiates by conducting a comprehensive examination of memory utilization repercussions on CDCs. Subsequently, it facilitates the dynamic clustering of physical machines and virtual machine deployments, ensuring a balanced utilization profile. Furthermore, it introduces the memory priority scheduling algorithm for CDC based on machine learning dynamic clustering algorithm (PMPD). Comparative evaluations against other algorithms underscore the prowess of PMPD in concurrently optimizing CPU and memory utilization, thereby minimizing the number of active PMs and diminishing energy consumption of CDCs.

关键词： Virtual machines Clustering algorithms Heuristic algorithms Resource management Energy consumption Scheduling algorithms Data centers Cloud computing Prediction algorithms machine learning algorithms Cloud data center (CDC) clustering algorithm machine learning virtual machine (VM)

来源：评论

学校读者我要写书评

暂无评论

Rethinking Discrepancy Analysis: Anomaly Detection via Meta-learning Powered Dual-Source Representation Differentiation

引用

IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING 2025年 22卷 8579-8592页

作者： Yao, Muyan Tao, Dan Qi, Peng Gao, Ruipeng Beijing Jiaotong Univ Sch Elect & Informat Engn Beijing 100040 Peoples R China China Elect Power Res Inst Beijing 100192 Peoples R China Beijing Jiaotong Univ Sch Cyberspace Sci & Technol Beijing 100040 Peoples R China

Industrial environments pose distinctive challenges for anomaly detection, primarily stemming from the complexities associated with high dimensionality and the dynamic nature of data patterns over time. These properties determine that the model's proper convergence on unlabeled data is unpromising, consequently leading to less efficient discrimination of anomalies in previous anomaly detection (AD) works. To address this problem, we present AnoDual, a novel, meta-learning AD framework. From the perspective of data reconstruction, we introduce the multi-memory enhanced VAE reconstructor M2ER, which learns to extract the most salient patterns in unlabeled noisy data through a self-supervised manner. This design eases impacts from potential anomalous components during data reconstruction, and enhances the discernibility of anomalies. To address performance degradation caused by the numerical deviation based AD scheme in most existing works, we design a dual-source self-supervised discriminator DSD, which examines characteristics in the domain of representations. This model actively assesses discrepancies between data pairs and representation pairs in parallel, and conducts AD on a fine-grained scale. In this way, anomalies that used to be unnoticed due to a less prominent numerical deviation can be spotted. Besides, we propose a meta-learning powered training pipeline to enable model training even when no real label is available, which is common in the industry. Extensive experiments on five large-scale real-world industrial datasets suggest that AnoDual achieves an average F1-Score with a substantial increment of 3.39 %, outperforming the latest state-of-the-art baseline. Note to Practitioners-A generative model plus a numerical threshold based detection approach currently takes a significant share in both academia and the industry. However, the performance of this workflow is not promising in actual applications, with multiple factors contributing to this situation.

关键词： Data models Training Numerical models Noise measurement Metalearning Anomaly detection machine learning algorithms Correlation Convergence Time series analysis meta-learning self-supervised learning industrial Internet of Things discrepancy analysis

来源：评论

学校读者我要写书评

暂无评论

Applying machine learning to Preselective Weighing of Moving Vehicles

引用

APPLIED SCIENCES-BASEL 2025年第4期15卷

作者： Kowaleczko, Pawel Kaminski, Tomasz Rychlicki, Mariusz Kasprzyk, Zbigniew Stawowy, Marek Trzeszkowski, Jacek Warsaw Univ Technol Inst Comp Sci Fac Elect & Informat Technol Pl Politech 1 St PL-00661 Warsaw Poland Rd & Bridge Res Inst Inst 1 St PL-03302 Warsaw Poland Warsaw Univ Technol Fac Transport Koszykowa 75 St PL-00662 Warsaw Poland Cypher Jacek Trzeszkowski Gen Jozefa Hallera 18 St PL-05091 Zabki Poland

The paper presents the general characteristics of weighing systems for vehicles in motion. A number of problems and constraints that accompany these systems to ensure adequate accuracy in the operation of these systems are pointed out. The efficient operation of WIM systems is also related to the proper preselection of vehicles for weighing in motion. The next part of the paper presents the basic classification and characteristics of machine learning algorithms, as well as examples of applications and implementations of these algorithms in various industries. The paper presents a model based on the XGBoost algorithm for estimating the weight of vehicles in motion, taking into account key characteristics of vehicles. The model was tested on large datasets from two locations in Poland, achieving high accuracy rates. The results indicate the model's potential in optimizing preselection systems, allowing for the effective identification of overloaded vehicles. Future work will focus on testing the model at other locations to verify its scalability and operational efficiency.

关键词： weigh-in-motion system machine learning algorithms preselective weighing of vehicles in motion

来源：评论

学校读者我要写书评

暂无评论

Non-Destructive Methods Based on machine learning for the Prediction of Sweet Potato Leaf Area: A Comparative Approach

引用

IEEE ACCESS 2025年 13卷 47389-47405页

作者： Ribeiro, Joao Everthon Da Silva Coelho, Ester Dos Santos da Silva, Antonio Gideilson Correia Henrique De Almeida Oliveira, Pablo da Silva, Elania Freire dos Santos, Gisele Lopes de Oliveira, Anna Kezia Soares Victor Lucas Lima, John Esfrain Pereira, Walter Maria Da Silveira, Lindomar Barros Jr, Aurelio Paes Fed Rural Semiarid Univ Ctr Agrarian Sci BR-59625900 Mossoro RN Brazil Univ Fed Paraiba Ctr Agr Sci BR-58397000 Areia PB Brazil

Leaf area is an essential parameter for studies of plant growth and physiology and is considered one of the main parameters for agricultural production. Leaf area determination methods are fundamental to understanding and predicting crop productivity. They can be classified as destructive or non-destructive. The study evaluated the performance of five methods for predicting the leaf area of sweet potato cultivars, including simple linear regression, artificial neural networks, support vector regression, adaptive neuro-fuzzy inference system (ANFIS), and random forest (RF). The coefficient of determination (R2), relative root mean squared error (RRMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and mean bias error (MBE) were used as criteria for choosing the best methods. The standardized Ranking Performance Index (sRPI) was used to classify the proposed methods. The ANFIS method performed better than the other methods analyzed (R $<^>{2} =0.8315$;RRMSE =0.0593;MAE =6.0789;MAPE =14.5741;MBE =0.000003;sRPI =1.00). Thus, the results indicated that the ANFIS method can be used as a non-destructive alternative for predicting the leaf area of sweet potato cultivars.

关键词： Mathematical models Biological system modeling Artificial intelligence Soil Predictive models Adaptation models Linear regression Accuracy Support vector machines Random forests Ipomoea batatas (L.) Lam machine learning algorithms non-destructive analysis precision agriculture leaf dimensions

来源：评论

学校读者我要写书评

暂无评论

Universal Adversarial Perturbations Against machine-learning-Based Intrusion Detection Systems in Industrial Internet of Things

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第2期12卷 1867-1889页

作者： Zhang, Sicong Xu, Yang Xie, Xiaoyao Guizhou Normal Univ Sch Cyber Sci & Technol Key Lab Informat & Comp Sci Guizhou Prov Guiyang 550001 Peoples R China

The security of the Industrial Internet of Things (IIoT) has emerged as a prominent concern in cyber-security due to the potential impact of attacks against IIoT on physical infrastructure. machine learning (ML)-based intrusion detection systems (IDSs) recently have been demonstrated to be an effective tool for protecting the systems in IIoT. However, the vulnerability of ML-based IDSs to adversarial attacks hinders their further application in IIoT. This article aims to further explore the adversarial attacks in IIoT to better evaluate the security of ML-based IDSs in this area. Our research primarily focuses on the generation of universal adversarial perturbations in IIoT, a topic that received limited attention in previous literature. Two novel attack methods based on a unified framework are proposed to utilize the original input-dependent adversarial perturbations of gradient-based or optimization-based adversarial attack methods to craft universal adversarial perturbations with better performance and transferability. The proposed attack methods conceal the underlying implementation details of the target attack methods, exploiting the original adversarial perturbations in a closed-box manner. This enhances their flexibility, making them applicable in a wider range of scenarios and enabling them to be combined with most gradient-based or optimization-based attack methods. Comprehensive experiments are conducted on three mainstream intrusion detection data sets, i.e., NSL-KDD, Gas Pipeline, and edge-IIoTset, to validate the effectiveness of the proposed methods. The preliminary experimental results demonstrate the feasibility of universal adversarial perturbations in IIoT and the superiority of the proposed methods to state-of-the-art attack methods.

关键词： Perturbation methods Intrusion detection Security Production Robustness machine learning algorithms Adversarial examples (AEs) industrial control systems (ICSs) Industrial Internet of Things (IIoT) intrusion detection universal adversarial perturbations universal adversarial perturbations

来源：评论

学校读者我要写书评

暂无评论

The role of social work in enhancing social governance: Policy, law, practice, and integration of machine learning for improved outcomes

引用

ALEXANDRIA ENGINEERING JOURNAL 2025年 118卷 208-215页

作者： Li, Xu Xue, Peng China Civil Affairs Univ Sch Social Work Beijing 102600 Peoples R China Hefei Univ Econ Sch Humanities & Law Hefei 230012 Anhui Peoples R China

Social workers need political and policy skills to promote social justice and systemic change. To build these abilities, one must understand the factors that drive political and policy engagement. This study employed mutual information to identify key factors affecting political and policy practice among social work students, focusing analysis on the most important factors. After feature selection, a refined dataset was subjected to machine learning models, such as the standard support vector machine (SVM), ordinal logistic regression (OLR), and ordinal random forest (ORF). The results indicate that the SVM obtains the highest accuracy of 91.38%, with OLR and ORF following closely behind at 91.00% and 89.63%, respectively. These findings indicate that the standard SVM is particularly well-suited for this dataset because of its capacity to establish optimal decision boundaries for the classification of political and policy engagement levels. Despite its somewhat lower accuracy, the ORF is a competitive option, while the OLR captures ordinal connections well using interpretability. Predictive modeling can improve teaching methodologies and focus interventions to encourage political and policy practice among social work students, according to the study. This project provides data-driven insights to educators and policymakers to improve student political and policy engagement.

关键词： Political practice Social work Feature selection machine learning algorithms Data-driven decision making

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：