检索结果-内蒙古大学图书馆

Multi-Party Federated Recommendation Based on Semi-Supervised learning

IEEE TRANSACTIONS ON BIG DATA 2024年第4期10卷 356-370页

作者： Liu, Xin Lv, Jiuluan Chen, Feng Wei, Qingjie He, Hangxuan Qian, Ying Chongqing Univ Posts & Telecommun Sch Software Engn Chongqing 400065 Peoples R China

Leveraging multi-party data to provide recommendations remains a challenge, particularly when the party in need of recommendation services possesses only positive samples while other parties just have unlabeled data. To address UDD-PU learning problem, this paper proposes an algorithm VFPU, Vertical Federated learning with Positive and Unlabeled data. VFPU conducts random sampling repeatedly from the multi-party unlabeled data, treating sampled data as negative ones. It hence forms multiple training datasets with balanced positive and negative samples, and multiple testing datasets with those unsampled data. For each training dataset, VFPU trains a base estimator adapted for the vertical federated learning framework iteratively. We use the trained base estimator to generate forecast scores for each sample in the testing dataset. Based on the sum of scores and their frequency of occurrence in the testing datasets, we calculate the probability of being positive for each unlabeled sample. Those with top probabilities are regarded as reliable positive samples. They are then added to the positive samples and subsequently removed from the unlabeled data. This process of sampling, training, and selecting positive samples is iterated repeatedly. Experimental results demonstrated that VFPU performed comparably to its non-federated counterparts and outperformed other federated semi-supervised learning methods.

关键词： Data models Federated learning Training Reliability Semisupervised learning Big Data machine learning algorithms Recommendation method vertical federated learning semi-supervised learning positive and unlabeled learning

来源：评论

学校读者我要写书评

暂无评论

Integrating machine learning for enhanced wildfire severity prediction: A study in the Upper Colorado River basin

引用

SCIENCE OF THE TOTAL ENVIRONMENT 2024年 952卷 175914页

作者： Han, Heechan Abitew, Tadesse A. Bazrkar, Hadi Park, Seonggyu Jeong, Jaehak Chosun Univ Dept Civil Engn Gwangju South Korea Texas A&M AgriLife Res Temple TX 76502 USA

Wildfires pose significant threats worldwide, requiring accurate prediction for mitigation. This study uses machine learning techniques to forecast wildfire severity in the Upper Colorado River basin. Datasets from 1984 to 2019 and key indicators like weather conditions and land use were employed. Random Forest outperformed Artificial Neural Network, achieving 72 % accuracy. Influential predictors include air temperature, vapor pressure deficit, NDVI, and fuel moisture. Solar radiation, SPEI, precipitation, and evapotranspiration also contribute significantly. Validation against actual severities from 2016 to 2019 showed mean prediction errors of 11.2 %, affirming the model's reliability. These results highlight the efficacy of machine learning in understanding wildfire severity, especially in vulnerable regions.

关键词： Artificial neural networks Landsat machine learning algorithms Normalized Burn Ratio Random Forest Upper Colorado River basin Wildfire severity prediction

来源：评论

学校读者我要写书评

暂无评论

Assessing the generalization of forecasting ability of machine learning and probabilistic models for complex climate characteristics

引用

STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT 2024年第8期38卷 2927-2947页

作者： Batool, Aamina Ali, Zulfiqar Mohsin, Muhammad Masmoudi, Atef Kartal, Veysi Satti, Samina Univ Punjab Coll Stat Sci Lahore Pakistan King Khalid Univ Coll Comp Sci Abha Saudi Arabia Siirt Univ Engn Fac Civil Engn Siirt Turkiye Univ Wah Dept Stat Rawalpindi Pakistan

Climate changes and global warming increase risk of recurrent extreme and complex climatic features. It necessitates accurate modeling and forecasting of climate phenomena for sustainable development goals. However, machine learning algorithms and advanced statistical models are extensively employed to analyze complex data and make predictions related to climate phenomena. It is important to have comprehensive knowledge to use these models and consider their potential implications. This study aims to evaluate and compare some popular machine learning and probabilistic methods by analyzing various time series indices associated with precipitation and temperature. For application, time series data of Standardized Precipitation Temperature Index (SPTI), Standardized Temperature Index (STI), Standardized Compound Drought and Heat Index (SCDHI), and Biased Diminished Weighted Regional Drought Index (BDWRDI) are used from various meteorological regions of Pakistan. The performance of each algorithm is compared using Residual Mean Square Error (RMSE) and Mean Average Error (MAE). The outcomes associated with this research indicate a higher preference of neural networks over machine learning methods in the training sets. However, the efficiency varies from model to model, indicator to indicator, time scale to time scale, and location to location during the testing phase. The most appropriate models are found by considering a list of candidates forecasting models and investigating the performance of each model.

关键词： Climate change machine learning algorithms Advanced statistical models Efficiency

来源：评论

学校读者我要写书评

暂无评论

machine learning and Deep learning for Loan Prediction in Banking: Exploring Ensemble Methods and Data Balancing

引用

IEEE ACCESS 2024年 12卷 193997-194019页

作者： Sayed, Eslam Hussein Alabrah, Amerah Rahouma, Kamel Hussein Zohaib, Muhammad Badry, Rasha M. Fayoum Univ Fac Comp & Informat Informat Syst Dept Faiyum Egypt Nahda Univ Fac Comp Sci Informat Syst Dept Bani Suwayf 62764 Egypt King Saud Univ Coll Comp & Informat Sci Dept Informat Syst Riyadh 11543 Saudi Arabia Minia Univ Fac Engn Elect Engn Dept Al Minya Egypt Lappeenranta Lahti Univ Technol Software Engn Dept Informat Syst Dept Lappeenranta 53851 Finland

The prediction of loan defaults is crucial for banks and financial institutions due to its impact on earnings, and it also plays a significant role in shaping credit scores. This task is a challenging one, and as the demand for loans increases, so does the number of applications. Traditional methods of checking eligibility are time-consuming and laborious, and they may not always accurately identify suitable loan recipients. As a result, some applicants may default on their loans, causing financial losses for banks. Artificial Intelligence, using machine learning and Deep learning techniques, can provide a more efficient solution. These techniques can use various classification algorithms to predict which applicants will likely be eligible for loans. This study uses five machine learning classification algorithms (Gaussian Naive Bayes, AdaBoost, Gradient Boosting, K Neighbors Classifier, Decision Trees, Random Forest, and Logistic Regression) and eight Deep learning algorithms (MLP, CNN, LSTM, Transformer, GRU, Autoencoder, ResNet, and DenseNet). The use of Ensemble Methods and SMOTE with SMOTE-TOMEK Techniques also has a positive impact on the results. Four metrics are used to evaluate the effectiveness of these algorithms: accuracy, precision, recall, and F1-measure. The study found that DenseNet and ResNet were the most accurate predictive models. These findings highlight the potential of predictive modeling in identifying credit disapproval among vulnerable consumers in a sea of loan applications.

关键词： Accuracy Random forests Predictive models Classification algorithms Prediction algorithms machine learning algorithms Logistic regression Support vector machines Ensemble learning Deep learning Customer loan prediction artificial intelligence data preprocessing model optimization machine learning deep learning classification models

来源：评论

学校读者我要写书评

暂无评论

System for Detection and Quantitative Evaluation of Compensatory Movement in Post-Stroke Patients Based on Wearable Sensor and machine learning Algorithm

引用

IEEE SENSORS JOURNAL 2024年第14期24卷 22830-22842页

作者： Ding, Kangjia Wang, Jiping Wang, Xiaojun Zhou, Linfu Xiong, Daxi Guo, Liquan Univ Sci & Technol China Sch Biomed Engn Suzhou Div Life Sci & Med Hefei 230052 Peoples R China Chinese Acad Sci Suzhou Inst Biomed Engn & Technol Suzhou 215163 Peoples R China Suzhou Xiangcheng Peoples Hosp Suzhou 215131 Peoples R China Nanjing Med Univ Affiliated Hosp 1 Dept Resp & Crit Care Med Nanjing 211166 Peoples R China

During upper limb rehabilitation training and assessment, compensatory movements can significantly impede the recovery of motor function in stroke survivors. Current methods for detecting these movements primarily rely on visual sensors, torso restraint belts, and pressure pads. However, these approaches present challenges such as obstacle occlusion, complex operation, and an inability to provide a quantitative assessment of compensatory movements. To address these issues, we developed a system for detecting and quantitatively evaluating compensatory movements, utilizing wearable sensors and machine learning algorithms. We recruited 30 healthy volunteers and 32 stroke patients to complete seven experimental motions. Their movement data were recorded using three inertial measurement units (IMUs) placed on the upper limbs and sternum. After preprocessing the data, six machine learning algorithms [K-nearest neighbor (KNN), support vector machine (SVM), multilayer perceptron (MLP), random forest (RF), eXtreme gradient boosting (XGBoost), and light gradient boosting machine (LightGBM)] were employed to train a multilabel classification model for detecting and quantitatively evaluating compensatory movements. The algorithm with the highest performance was selected as the system's classifier. The results show that the proposed system improves the number and accuracy of compensatory movement detection and achieves accurate classification and detection of seven compensatory movements (average accuracy is 95.7%, F1-score is 96.8%). Furthermore, the KNN classifier classified compensatory movements as mild, moderate, and severe according to the different severity of the compensatory movements and obtained an average accuracy of 93.7%, which proved the reliability and responsiveness of the system. This system is applicable to stroke patients in clinical, home, and community-based rehabilitation settings, enabling a quantitative assessment of compensatory movements.

关键词： Stroke (medical condition) Sensors Shoulder Motors Data models Compensation detection machine learning algorithms quantitative evaluation wearable sensors quantitative evaluation wearable sensors

来源：评论

学校读者我要写书评

暂无评论

machine learning models for predicting hospitalization and mortality risks of COVID-19 patients

引用

EXPERT SYSTEMS WITH APPLICATIONS 2024年 240卷

作者： de Holanda, Wallace Duarte Chaves e Silva, Lenardo Sobrinho, alvaro Alvares de Carvalho Cesar Fed Univ Semiarid Francisco Mota 572 BR-59625900 Mossoro RN Brazil Fed Univ Agreste Pernambuco BR-55292270 Garanhuns PE Brazil

Even though the World Health Organization declared the end of the pandemic, COVID-19 is still considered an endemic disease, affecting many people worldwide. As a result, hospitalization rates can continue to strain healthcare systems, particularly in low and middle-income countries. Thus, this article presents two predictive machine learning (ML) models to identify COVID-19 patients more likely to be hospitalized or face a higher mortality risk, aiming to assist physicians and healthcare administrators in making informed decisions. Previous research primarily focused on using expensive input data, such as clinical examinations and images, to implement ML models. In contrast, we implemented the models using information on patients' demographics, vaccination history, symptoms, and underlying health conditions. We trained the ML algorithms using data from the OpenDataSus database, focusing on patients undergoing hospital treatment in Sao Paulo, Brazil. Our study tested 14 different ML algorithms to determine their effectiveness: gradient boosting, light gradient boosting machine, logistic regression, linear discriminant analysis, ridge, AdaBoost, random forest, support vector machine, k-nearest neighbor, extra tree, decision tree, naive Bayes, quadratic discriminant analysis, and dummy. The results showed that the gradient-boosting model outperformed the others, achieving an accuracy rate of 83% and an Area Under the Curve (AUC) of 0.89 for predicting mortality risk. Moreover, the model achieved an accuracy rate of 71% and an AUC of 0.75 for predicting hospitalization risk. Our analysis evidence that including information on the number of vaccine doses can enhance the prediction results for hospitalization and mortality of patients. Based on these significant findings, we have developed a web application to assist healthcare professionals in identifying patients at a higher risk of severe outcomes from COVID-19.

关键词： Coronavirus disease 2019 Risk of worsening machine learning algorithms Predictive model

来源：评论

学校读者我要写书评

暂无评论

Electronic Skin: Opportunities and Challenges in Convergence with machine learning

引用

ANNUAL REVIEW OF BIOMEDICAL ENGINEERING 2024年第1期26卷 331-355页

作者： Koo, Ja Hoon Lee, Young Joong Kim, Hye Jin Matusik, Wojciech Kim, Dae-Hyeong Jeong, Hyoyoung Sejong Univ Dept Semicond Syst Engn Seoul South Korea Sejong Univ Inst Semicond & Syst IC Seoul South Korea MIT Comp Sci & Artificial Intelligence Lab Cambridge MA USA MIT Dept Elect Engn & Comp Sci Cambridge MA USA Inst for Basic Sci Korea Ctr Nanoparticle Res Seoul South Korea Seoul Natl Univ Inst Chem Proc Sch Chem & Biol Engn Seoul South Korea Seoul Natl Univ Dept Mat Sci & Engn Seoul South Korea Seoul Natl Univ Interdisciplinary Program Bioengn Seoul South Korea Univ Calif Davis Dept Elect & Comp Engn Davis CA USA

Recent advancements in soft electronic skin (e-skin) have led to the development of human-like devices that reproduce the skin's functions and physical attributes. These devices are being explored for applications in robotic prostheses as well as for collecting biopotentials for disease diagnosis and treatment, as exemplified by biomedical e-skins. More recently, machine learning (ML) has been utilized to enhance device control accuracy and data processing efficiency. The convergence of e-skin technologies with ML is promoting their translation into clinical practice, especially in healthcare. This review highlights the latest developments in ML-reinforced e-skin devices for robotic prostheses and biomedical instrumentations. We first describe technological break-throughs in state-of-the-art e-skin devices, emphasizing technologies that achieve skin-like properties. We then introduce ML methods adopted for control optimization and pattern recognition, followed by practical applications that converge the two technologies. Lastly, we briefly discuss the challenges this interdisciplinary research encounters in its clinical and industrial transition.

关键词： electronic skins machine learning algorithms robotic skins electronic prostheses wearable electronics skin-mounted electronics

来源：评论

学校读者我要写书评

暂无评论

Wireless Edge Caching and Content Popularity Prediction Using machine learning

引用

IEEE CONSUMER ELECTRONICS MAGAZINE 2024年第4期13卷 32-41页

作者： Krishnendu, S. Bharath, B. N. Bhatia, Vimal Nebhen, Jamel Dobrovolny, Michal Ratnarajah, Tharmalingam Indian Inst Technol Indore Dept Elect Engn Indore India Indian Inst Technol Dharwad Dept Elect Engn Dharwad India Indian Inst Technol Indore Indore India Univ Hradec Kralove Hradec Kralove Czech Republic Prince Sattam Bin Abdulaziz Univ Alkharj Saudi Arabia Univ Hradec Kralove Appl Informat Hradec Kralove Czech Republic Univ Edinburgh Inst Digital Commun Edinburgh Midlothian Scotland

The ever pervasive growth in information services and technology has resulted in the outbreak of demand for data in the wireless networks. This has made the network operators to ponder over the imminent difficulties, such as computing capabilities and fronthaul-backhaul link capacities. Hence, to bridge the gap between the cloud capacity and requirement of mobile services by the network edges, edge computing, and caching techniques have been gaining more and more attention from researchers across the world. Further, motivated by the successful applications of machine learning (ML) in solving complex and dynamic problems, in this article, it has been used to advance edge caching capabilities. The proposed ML-based algorithms have been evaluated and proved to have better performance compared with the existing conventional algorithms. The mean squared error (mse), for the proposed deep learning (DL) algorithm, is 20-times less than the existing algorithms and while comparing with the simple neural network, the gain in mse for the proposed DL algorithm is observed around 27%. Similarly for the federated learning-based caching algorithm, the average cache hit gain is of the order of $10<<^>>algorithms$, hence demonstrating the benefit of the proposed algorithms. In addition, opportunities and challenges for a promising upcoming future of ML in edge computing and content popularity prediction have also been discussed.

关键词： Q-learning Wireless networks machine learning algorithms Device-to-device communication Prediction algorithms Markov processes Consumer electronics

来源：评论

学校读者我要写书评

暂无评论

Using machine learning to improve the estimate of US background ozone

引用

ATMOSPHERIC ENVIRONMENT 2024年 316卷

作者： Hosseinpour, Farnaz Kumar, Naresh Tran, Trang Knipping, Eladio Desert Res Inst Reno NV 89512 USA Univ Nevada Reno Reno NV USA Ramboll Novato CA 94945 USA Elect Power Res Inst Palo Alto CA 94304 USA

U.S. background ozone is defined as ground-level ozone in the absence of any domestic anthropogenic emissions. It plays an important role in the U.S. regulatory framework, as it corresponds to the theoretical minimum ozone that can be achieved at a location. Although some studies estimated background ozone from observations, it cannot be measured directly, and therefore air quality models have been used to estimate its levels across different parts of the country. These estimates are uncertain given the inherent uncertainty in air quality modeling. Recently data fusion techniques have been used to improve those estimates using ground-level ozone observations and satellite data. Here, we used two adjustment methods to improve on air quality model-derived estimates of background ozone: a Multi-Variate Linear Regression (MVLR) model and a machine learning (ML) algorithm, called Random Forest (RF), with a focus on thirteen urban areas in the U.S. These adjustments regress observed ozone on the simulated background and anthropogenic ozone (by a regional air quality model) and their combinations with spatial and temporal variables. The RF-ML algorithm showed the biggest improvement in model performance, compared to MVLR and the base air quality model predictions. Applying the MVLR adjustment, the background ozone estimates increased in most locations. Furthermore, applying the RF ML model adjustment increased background ozone estimates even more, highlighting the enhanced contribution of background ozone in recent years that constitutes a larger portion of total ozone and exceeds NAAQS in some regions. Our results are consistent with other observational and modeling studies in that the highest background ozone levels were estimated in the western U.S. and over higher elevations.

关键词： US background ozone Anthropogenic ozone Air quality modeling machine learning algorithms Data fusion Air pollution

来源：评论

学校读者我要写书评

暂无评论

A Multi-Classification Accessment Framework for Reproducible Evaluation of Multimodal learning in Alzheimer's Disease

引用

IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS 2024年第4期21卷 559-572页

作者： Nan, Fengtao Li, Shunbao Wang, Jiayu Tang, Yahui Qi, Jun Zhou, Menghui Zhao, Zhong Yang, Yun Yang, Po Yunnan Univ Sch informat Kunming Peoples R China Univ Sheffield Dept Comp Sci Sheffield S1 4DP England Yunnan Univ Natl Pilot Sch Software Kunming 650500 Peoples R China Xian Jiao Tong Liverpool Univ Dept Comp Suzhou 215123 Peoples R China Yunnan Prov Dept Neurol Peoples Hosp 1 Kunming 650500 Peoples R China

Multimodal learning is widely used in automated early diagnosis of Alzheimer's disease. However, the current studies are based on an assumption that different modalities can provide more complementary information to help classify the samples from the public dataset Alzheimer's Disease Neuroimaging Initiative (ADNI). In addition, the combination of modalities and different tasks are external factors that affect the performance of multimodal learning. Above all, we summrise three main problems in the early diagnosis of Alzheimer's disease: (i) unimodal vs multimodal;(ii) different combinations of modalities;(iii) classification of different tasks. In this paper, to experimentally verify these three problems, a novel and reproducible multi-classification framework for Alzheimer's disease early automatic diagnosis is proposed to evaluate and verify the above issues. The multi-classification framework contains four layers, two types of feature representation methods, and two types of models to verify these three issues. At the same time, our framework is extensible, that is, it is compatible with new modalities generated by new technologies. Following that, a series of experiments based on the ADNI-1 dataset are conducted and some possible explanations for the early diagnosis of Alzheimer's disease are obtained through multimodal learning. Experimental results show that SNP has the highest accuracy rate of 57.09% in the early diagnosis of Alzheimer's disease. In the modality combination, the addition of Single Nucleotide Polymorphism modality improves the multi-modal machine learning performance by 3% to 7%. Furthermore, we analyse and discuss the most related Region of Interest and Single Nucleotide Polymorphism features of different modalities.

关键词： Magnetic resonance imaging Medical diagnostic imaging Alzheimer's disease Feature extraction Task analysis machine learning machine learning algorithms Multi-modal learning multi-modality data alzheimer's disease

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：