A5100366721- Data Stream Mining Techniques
- Machine Learning and Data Classification
- E-commerce and Technology Innovations
- Advanced Computational Techniques and Applications
- Structural Load-Bearing Analysis
- Topic Modeling
- Structural Behavior of Reinforced Concrete
- Geotechnical Engineering and Underground Structures
- X-ray Diffraction in Crystallography
- Time Series Analysis and Forecasting
- Anomaly Detection Techniques and Applications
- Advanced Bandit Algorithms Research
- Spectroscopy and Chemometric Analyses
- Semantic Web and Ontologies
- Collaboration in agile enterprises
- Piezoelectric Actuators and Control
- Digital Marketing and Social Media
- Air Quality Monitoring and Forecasting
- Technology and Security Systems
- Metallurgical and Alloy Processes
- Imbalanced Data Classification Techniques
- Advanced Decision-Making Techniques
- Geoscience and Mining Technology
- Thermal Expansion and Ionic Conductivity
- Natural Language Processing Techniques
Shanghai University
2009-2025
University of Technology Sydney
2020-2024
University of Science and Technology of China
2024
Anhui University of Science and Technology
2014-2022
Los Alamos National Laboratory
2022
Nanjing University of Posts and Telecommunications
2016
Chinese Academy of Sciences
2015
South China Normal University
2013
Taiyuan University of Science and Technology
2012
Southwest University of Science and Technology
2010
Abstract Interlayer engineering is a promising strategy to modify the structure of layered vanadium‐based oxides with optimized ion‐diffusion capability, during which role interlayer crystal water in tuning charge storage properties should be clarified. Herein, series hydrated V 2 O 5 · n H xerogels varying contents ( ) obtained by differentiating temperature parameter. Results show that value properly modified best not too large or small, is, when equals 0.26, electrode exhibits discharge...
The captured underwater images suffer from color cast and haze effect caused by absorption scattering. These interdependent phenomena jointly degrade images, resulting in failure of autonomous machines to recognize image contents. Most existing learning-based methods for enhancement (UIE) treat the degraded process as a whole ignore interaction between correction dehazing. Thus, they often obtain unnatural results. To this end, we propose novel joint network optimize results dehazing...
Traditional recommender systems usually take the user-platform paradigm, where users are directly exposed under control of platform's recommendation algorithms. However, defect algorithms may put in very vulnerable positions this paradigm. First, many sophisticated models often designed with commercial objectives mind, focusing on benefits, which hinder their ability to protect and capture users' true interests. Second, these typically optimized using data from all users, overlook individual...
In this work, several thermodynamic assessments adopted widely for ZrO 2 –CaO system are reviewed and examined, the existing discrepancies summarized, a new assessment is carried out based on formation enthalpy of two compounds (CaZr 4 O 9 Ca 6 Zr 19 44 ) experimental activity data concerning cubic solid solution. The parameters all phases have been optimized by least squares minimization procedure, self consistent set Gibbs energy has derived, which can be safely used to extrapolate into...
In nonstationary environments, data distributions can change over time. This phenomenon is known as concept drift, and the related models need to adapt if they are remain accurate. With gradient boosting (GB) ensemble models, selecting which weak learners keep/prune maintain model accuracy under drift nontrivial research. Unlike existing such AdaBoost, directly compare learners' performance by their (a metric between [0, 1]), in GB, measured with different scales. To address measurement...
Concept drift arises from the uncertainty of data distribution over time and is common in stream. While numerous methods have been developed to assist machine learning models adapting such changeable data, problem improperly keeping or discarding samples remains. This may results loss valuable knowledge that could be utilized subsequent points, ultimately affecting model's accuracy. To address this issue, a novel method called segmentation-based stream (TS-DM) help segment learn streaming...
Particle breakage in sand significantly influences the stress and deformation predictions of spherical cavity expansion theories. However, existing theories often overlook or lack a quantitative description particle breakage, leading to significant errors actual results. Based on Simple Critical State Sand (SIMSAND) model, considering effect, traditional Euler's problem drainage is converted into set first-order ordinary differential equations described by Lagrangian. Fontainebleau taken as...
Uncertain changes in data streams present challenges for machine learning models to dynamically adapt and uphold performance real-time. Particularly, classification boundary change, also known as real concept drift, is the major cause of deterioration. However, accurately detecting drift remains challenging because theoretical foundations existing detection methods - two-sample distribution tests monitoring error rate, both suffer from inherent limitations such inability distinguish virtual...
The purpose of this study was to use the stochastic simulation and estimation method evaluate effects sample size number samples per individual on model development evaluation. pharmacokinetic parameters inter- intra-individual variation were obtained from a population clinical trials amlodipine. Stochastic performed efficiencies different sparse sampling scenarios estimate compartment model. Simulated data generated 1000 times three candidate models used fit sets. Fifty-five kinds...
Piezoelectric inchworm actuators have a wide application in the field of Nano positioning and ultra-precision detecting instruments which depend on characteristics large stroke, high resolution rigidity, quick response speed, small size, driving force, low power consumption, not being affected by electromagnetic interference, so on. A new piezoelectric actuator based principle flexible amplification is developed this paper. In moving mechanism actuator, its clamping adopts symmetrical lever...
With the proposal of goal “carbon peak, carbon neutrality”, concept environmental protection has become increasingly popular. To explore characteristics and influencing factors consumers’ green consumption behavior against a dual-carbon background, this study proposed research variables hypotheses about factors. purpose, variables, hypotheses, questionnaire was designed. Consumers in Anhui Province were chosen as samples. SPSS26.0 employed to conduct reliability validity analysis,...
The advances of network technology and mobile communication are making eHealth possible. In systems, physiological data relevant context-aware acquired continuously in real time. At the same time, such large-scale results huge challenges aspect real-time big processing since appears form stream. Therefore, we propose a novel incremental learning algorithm, namely α-SVMSGD, which improves SVMSGD (Support Vector Machine-Stochastic Gradient Descent) algorithm by updating training with...