Search

Purpose: Patients with dementia symptoms often struggle to limit movements during PET examinations, necessitating motion compensation in brain PET imaging to ensure the high image quality needed for diagnostic accuracy. This study validates a data-driven motion-compensated (MoCo) PET brain image reconstruction algorithm that corrects head motion by integrating the detected motion frames and their associated rigid body transformations into the iterative image reconstruction. Validation was conducted using phantom scans, healthy volunteers, and clinical patients using four radiotracers with distinct tracer activity distributions. Methods: We conducted technical validation experiments of the algorithm using Hoffman brain phantom scans during a series of controlled movements, followed by two blinded reader studies assessing image quality between standard images and MoCo images in 38 clinical patients receiving dementia scans with [18F]Fluorodeoxyglucose, [18F]N-(3-iodopro-2E-enyl)-2beta-carbomethoxy-3beta-(4'-methylphenyl)-nortropane, [18F]flutemetamol, and a research group comprising 25 elderly subjects scanned with [18F]fluoroethoxybenzovesamicol. Results: The Hoffman brain phantom study demonstrated the algorithm's capability to detect and correct for even minimal movements, 1-mm translations and 1⁰ rotations, applied to the phantom. Within the clinical cohort, where standard images were deemed suboptimal or non-diagnostic, all MoCo images were classified as having acceptable diagnostic quality. In the research cohort, MoCo images consistently matched or surpassed the standard image quality even in cases with minimal head movement, and the MoCo algorithm never led to degraded image quality. Conclusion: The PET brain MoCo reconstruction algorithm was robust and worked well for four different tracers with markedly different uptake patterns. Moco images markedly improved the image quality for patients who were unable to lie still during a PET examination and obviated the need for any repeat scans. Thus, the method was clinically feasible and has the potential for improving diagnostic accuracy. [ABSTRACT FROM AUTHOR]

In order to smooth the impact of distributed photovoltaic and load fluctuations on the distribution network, this paper proposes a dual time scale distribution network reactive voltage control method based on data-driven and source-load uncertainty. The proposed method can be divided into two stages, the first stage is centralized optimization, which performs long time scale scheduling of slow regulation devices in active distribution networks, and converts the original centralized optimization problem into a second-order cone mixed-integer planning problem after second-order cone and linearization theory. The second stage is based on data-driven distributed control, which is based on the information communication between the regional controllers to achieve short time scale optimization control of photovoltaic inverters, where the regional controllers are modeled based on deep neural networks. In order to make the model training more effective, the paper uses an improved generative adversarial network to generate a large number of source-load samples conforming to the real distribution for the training of the regional controllers. Through the analysis of the improved IEEE33 model, the optimal control of reactive voltage with dual time scales is achieved, which keeps the voltage within the safety range while significantly reducing the network loss, and the computation time of the regional controller based on the data-driven model is reduced by 96.52% compared to the centralized control, which makes the frequent control possible in the face of more complex power systems. [ABSTRACT FROM AUTHOR]

Landslide, as a significant global natural hazard, threatening human settlements and the natural environment. The present study introduces a novel approach to landslide susceptibility assessment by integrating the Forest Attribute Penalty (FPA) model with three ensemble algorithms—AdaBoost (AB), Rotation Forest (RF), and Random Subspace (RS)—and utilizing the Evidential Belief Function (EBF) to weight the classes of landslide-related factors. To evaluate the performance of the developed methodology, Yanchuan County, China, was chosen as the appropriate study area. Three hundred and eleven landslide areas were identified through remote sensing and field investigations, which were randomly divided into 70% for model training and 30% for model evaluation, whereas sixteen landslide – related factors were considered, such as elevation, slope aspect, profile curvature, plan curvature, convergence index, slope length, terrain ruggedness index, topographic position index, distance to roads, distance to rivers, NDVI, land use, soil, rainfall, and lithology. EBF was employed to analyze the spatial correlation between these factors and landslide occurrences, providing the class weights of each factor for the implementation of FPA and the ensemble models. The next step involved the generation of the landslide susceptibility maps based on the models, with findings showing that more than half of the study area is classified as very low susceptibility. Model performance was assessed using receiver operating characteristic (ROC) curves and other statistical metrics, with the RFFPA model achieving the highest predictive ability, with AUC values of 0.878 and 0.890 for training and validation datasets, respectively. The AFPA and RSFPA hybrid models, however, demonstrated weaker predictive abilities compared to the FPA model. The study highlights the importance of optimizing model performance and evaluating the suitability of ensemble approaches, emphasizing the role of topographical and environmental settings in influencing model accuracy. The use of EBF for weight calculation proved crucial in improving model outcomes, suggesting that this approach could be further refined and adapted to other regions with similar geomorphological settings for better land use planning and risk management. [ABSTRACT FROM AUTHOR]

Rapid parking trajectory planning is very import for intelligent cyber-physical transportation systems. A three-stage Gaussian pseudo-spectrum data-driven fuzzy logic parallel parking trajectory planning algorithm is proposed in this work. First, a three-stage parking trajectory planning dynamic optimization problem (DOP) is established based on the three stages parking division and vehicle kinematics model. Subsequently, an improved three-stage Gaussian pseudo-spectrum method (T-GPM) is proposed to solve the typical DOPs so as to construct the database. With analyzing the database, fuzzy rules are designed to achieve fuzzy logic-based data-driven parking trajectory planning, which would extremely decrease the planning time of parking trajectory. To verify the performance of the proposed method, numerical simulation tests are conducted and results show that with using the data-driven fuzzy logic strategy, the proposed planning method can quickly and effectively generate parking trajectories online with small deviation, and the calculation time is reduced by over 99% compared to T-GPM. Furthermore, the CarSim co-simulation tests and real car parking tests are conducted with different scenarios. Experiments demonstrate that feasible parking trajectories can be generated by the proposed method and the trajectories are tracked with no collision, revealing the effectiveness of the improvement. [ABSTRACT FROM AUTHOR]

False data injection attacks (FDIAs) targeting AC state estimation pose significant challenges, especially when only power measurements are available, and voltage measurements are absent. Current machine learning-based approaches struggle to effectively control state estimation errors and are confined to the data distribution of training sets. To address these limitations, we propose the physics-informed extrapolative adversarial variational autoencoder (PI-ExAVAE) for generating controllable and stealthy false data injections. By incorporating physically consistent priors derived from the AC power flow equations, which enforce both the physical laws of power systems and the stealth requirements to evade bad data detection mechanisms, the model learns to generate attack vectors that are physically plausible and stealthy while inducing significant and controllable deviations in state estimation. Experimental results on IEEE-14 and IEEE-118 systems show that the model achieves a 90% success rate in bypassing detection tests for most attack configurations and outperforms methods like SAGAN by generating smoother, more realistic deviations. Furthermore, the use of physical priors enables the model to extrapolate beyond the training data distribution, effectively targeting unseen operational scenarios. This highlights the importance of integrating physics knowledge into data-driven approaches to enhance adaptability and robustness against evolving detection mechanisms. [ABSTRACT FROM AUTHOR]

Precise temperature and heating power control are crucial for crystal quality and production efficiency in the Czochralski single-crystal growth process. Existing sensor technologies can only monitor these parameters in real time, lacking the ability to predict future trends, which limits the ability to implement preventive control before issues arise. To address this, a temperature and heating power prediction model based on Long Short-Term Memory (LSTM) is proposed and developed using extensive production data. Spearman's rank correlation coefficient is applied to identify the key parameters related to temperature and heating power. Hyperparameter optimization uses Particle Swarm Optimization (PSO) to improve prediction accuracy. The performance of the PSO-LSTM model is compared with two other widely used prediction models, demonstrating its superior predictive capability. The results show that the PSO-LSTM model achieves highly accurate temperature and heating power predictions in the crystal growth process, with a Mean Absolute Error (MAE) of 0.0295 for temperature and 0.0392 for heating power, further validating its effectiveness for real-time predictive control. [ABSTRACT FROM AUTHOR]

The traditional power load forecasting learning method has problems such as overfitting and incomplete learning of time series information when dealing with complex nonlinear data, which affects the accuracy of short–medium term power load forecasting. A joint learning method, LSVM-MKL, was proposed based on the bidirectional promotion of deep kernel learning (DKL) and multiple kernel learning (MKL). The multi-kernel method was combined with the input layer, the highest coding layer, and the highest encoding layer to model the network of the stack autoencoder (SAE) to obtain more comprehensive information. At the same time, the deep kernel was integrated into the optimization training of Gaussian multi-kernel by means of the nonlinear product to form the nonlinear composite kernel. Through a large number of reference datasets and actual industrial data experiments, it was shown that compared with the Elman and LSTM-Seq2Seq methods, the proposed method achieved a higher prediction accuracy of 4.32%, which verified its adaptability to complex time-varying power load forecasting processes and greatly improved the accuracy of power load forecasting. [ABSTRACT FROM AUTHOR]

Around 2.6 billion people are at risk of tooth carries and fluorosis worldwide. Quetta is the worst affected district in Balochistan plateau. Endemic abnormal groundwater fluoride ( F - ) lacks spatiotemporal studies. This research integrates geospatial distribution, geochemical signatures, and data driven method for evaluating F - levels and population at risk. Groundwater F - ranged from 0 to 3.4 mg/l in (n = 100) with 52% samples found unfit for drinking. Through geospatial IDW tool hotspot areas affected with low and high groundwater F - levels were identified. Geochemical distribution in geological setups recognized sediment variation leads to high F - (NaHCO3) and low F - (CaHCO3) water types in low elevation (central plain) and high elevation (mountain foot) respectively. Results of the modified water quality index identified 60% samples to be unsuitable for drinking. Support vector machine (SVM), random forest regression (RFR) and classification and regression tree (CART) machine learning models found Na + , Salinity and Ca + 2 as important contributing variables in groundwater F - prediction. CART model with R2 value of 0.732 outperformed RFR and SVM in predicting F - . Noncarcinogenic health risk vulnerability from F - increased from Adults < Teens < Children < Infants. Infants and children with hazard quotient values of 11.3 and 4.2 were the most vulnerable population at risk for consuming F - contaminated groundwater. The research emphasizes on both nutritional need and hazardous effect of F - , and development of desirable limit for F - . [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Social security event prediction can provide critical early warnings and support for public policies and crisis responses. The rapid development of communication networks has provided a massive data analysis base, including social media, economic data, and historical event records, for social security event prediction based on data-driven approaches. The advent of data-driven approaches has revolutionized the prediction of these events, offering new theoretical insights and practical applications. Aiming at offering a systematic review of current data-driven prediction methods used in social security, this paper delves into the progress of this research from three novel perspectives, prediction factors, technical methods, and interpretability, and then analyzes future development trends. This paper contributes key insights into how social security event prediction can be improved and hopefully offers a comprehensive analysis that goes beyond the existing literature. [ABSTRACT FROM AUTHOR]

A key challenge in understanding the functional organization of the visual cortex stems from the fact that only a small proportion of the objects experienced during natural viewing can be presented in a typical experiment. This constraint often leads to experimental designs that compare responses to objects from experimenter-defined stimulus conditions, potentially limiting the interpretation of the data. To overcome this issue, we used images from the THINGS initiative, which provides a systematic sampling of natural objects. A data-driven analysis was then applied to reveal the functional organization of the visual brain, incorporating both perceptual and neural responses to these objects. Perceptual properties of the objects were taken from an analysis of similarity judgments, and neural properties were taken from whole-brain fMRI responses to the same objects. Partial least squares regression (PLSR) was then used to predict neural responses across the brain from the perceptual properties while simultaneously applying dimensionality reduction. The PLSR model accurately predicted neural responses across the visual cortex using only a small number of components. These components revealed smooth, graded neural topographies, which were similar in both hemispheres and captured a variety of object properties including animacy, real-world size, and object category. However, they did not accord in any simple way with previous theoretical perspectives on object perception. Instead, our findings suggest that the visual cortex encodes information in a statistically efficient manner, reflecting natural variability among objects. [ABSTRACT FROM AUTHOR]

This study demonstrates that machine learning from seismograms, obtained from commonly deployed seismometers, can identify the early stages of slope failure in the field. Landslide hazards negatively impact the economy and public through disruption, damage of infrastructure and even loss of life. Triggering factors leading to landslides are broadly understood, typically associated with rainfall, geological conditions and steep topography. However, early warning at slope scale of an imminent landslide is more challenging. Through semi-supervised learning for seismic event detection from continuous seismic recordings over a period of about 10 years, we demonstrate that timely landslide induced displacement prediction is possible, providing the basis for landslide early warning systems. Our proposed methodology detects and characterises seismic precursors to landslide events making use of seismic recordings near an active slow moving earth slide-flow using a semi-supervised Siamese network. This data driven methodology identifies increase in microseismicity, and the change in the frequency spectrum of that microseismicity which identify key stages prior to a failure: 'rest', 'precursor' and 'active'. Due to the semi-supervised nature of Siamese networks, the methodology is adaptable to discovering new types of distinct events, making it an ideal solution for precursor detection at new sites. [ABSTRACT FROM AUTHOR]

Background: Neurocognitive and neuropsychiatric symptoms are essential clinical manifestations of age-related cognitive impairment, yet their patterns of co-existence remain unclear through the cognitive continuum. Objective: To examine the associations of person-centered cluster-derived patterns, based on a comprehensive collection of domain-specific cognitive and neuropsychiatric assessments, with neuroimaging markers and dementia risk. Methods: 641 participants were included in the analysis from memory clinics in Singapore. Latent class analysis was applied to define clusters of individuals with different clinical patterns. The associations between identified clinical groups with neuroimaging markers of cerebrovascular diseases and neurodegeneration were analyzed using logistic regression models. Cox proportional hazard models were applied for incident dementia. Results: Three latent classes differing in neurocognitive and neuropsychiatric impairment were identified (Class 1 "memory impairment only"; Class 2 "global cognitive impairment"; Class 3 "global cognitive and neuropsychiatric impairment"). Compared with Class 1, Class 2 and 3 were associated with smaller brain volumes, moderate-to-severe cortical atrophy and medial temporal lobe atrophy, and the presence of all cerebrovascular lesions. Moreover, compared with Class 2, Class 3 had smaller brain volumes, moderate-to-severe cortical atrophy and presence of intracranial stenosis. Additionally, compared to Class 1, Class 2 (hazard ratio [HR] = 3.84, 95%CI 2.11–7.00), and Class 3 (HR = 6.92, 95%CI 2.84–16.83) showed an increased risk of incident dementia. Conclusions: Participants characterized by multi-domain cognitive impairment and co-occurrence of cognitive and neuropsychiatric impairment showed the highest risk of incident dementia, which may be attributed to both neurodegenerative and cerebrovascular pathologies. [ABSTRACT FROM AUTHOR]

Virtual power plant needs to use advanced coordinated control technology to aggregate a large amount of new energy to reliably meet the regulatory needs of the superior power grid. Currently, virtual power plant aggregation technology considering reliability effectively alleviates the problems of low reliability of traditional virtual power plants and poor absorption capacity of new energy. However, in the process of solving the optimization scheme, the traditional optimization solution based on physical models is faced with great challenges due to the complex characteristics such as diversity and heterogeneity of virtual power plant aggregation models. Therefore, a data-driven virtual power plant aggregation method is proposed. The dispatching characteristics of different virtual power plant clusters can be effectively expressed by using the load data, the historical dispatching data of virtual power plant clusters and the data-driven technology. The packaging model reflecting the reliability difference of virtual power plant assemblies is established. The results show that the calculation results indicate that the root mean square error of the model is only 0.2134. Compared to LSTM training model and BP neural networks, the RMSE has decreased by 44.22% and 54.41%, respectively, while the MAE has decreased by 48.32% and 57.84%, respectively. This method has good accuracy. At the same time, this method provides a new method for complex and heterogeneous power system dispatching operation of China's new power system. [ABSTRACT FROM AUTHOR]

The prediction of debris flows is essential for safeguarding infrastructure and minimizing the economic losses associated with the hazards. Traditional empirical and theoretical models, while providing foundational insights, often struggle to capture the complex and nonlinear behaviors inherent in debris flows. This study aims to enhance debris flow prediction by integrating theoretical modeling with data-driven approaches. We model debris flow as a viscoplastic fluid, employing the Herschel–Bulkley rheological model to describe its behavior. By combining the kinematic wave model with lubrication theory, we develop a comprehensive theoretical framework that encapsulates the mechanical physics of debris flows and identifies key governing parameters. Numerical solutions of this theoretical model are utilized to generate an extensive training dataset, which is subsequently used to train a support vector regression (SVR) model. The SVR model targets slide depth and velocity upon impact, using explanatory variables including yield stress, material density, source area depth and length, and slope length. The model demonstrates high predictive accuracy, achieving coefficients of determination R 2 of 0.956 for slide depth and 0.911 for slide velocity at impact. Additionally, the relative residuals σ are primarily distributed within the range of −0.05 to 0.05 for both slide depth and slide velocity upon impact. These results indicate that the proposed hybrid model not only incorporates the fundamental physical mechanisms governing debris flows but also significantly enhances predictive performance through data-driven optimization. This study underscores the critical advantage of merging physical models with machine learning techniques, offering a robust tool for improved debris flow prediction and risk assessment, which can inform the development of more effective early warning systems and mitigation measures. [ABSTRACT FROM AUTHOR]

Copyright of Computer Measurement & Control is the property of Magazine Agency of Computer Measurement & Control and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

ObjectiveAiming at the problems of unknown ship model parameters and external disturbance and servo constraints, this paper proposes a method for the data-driven online identification of ship parameters and iterative analytical calculation of the optimal control quantity of track tracking control. MethodA three degrees of freedom dynamics equation of a double propeller ship is constructed, and the extended state observer-multiple innovation recursive least squares interactive algorithm is designed by collecting the motion data of the ship. By approximating the identified ship motion model to a time-invariant linear model in the sampling period, the ship trajectory tracking problem can be transformed into a linear quadratic optimization control problem with constraints and disturbances. The weighted matrix and penalty function are introduced to construct the quadratic performance index including trajectory error, external disturbance, and control constraint inequality. The precise integration method is then used to obtain the analytical solution of the matrix Riccati differential equation and the iterative calculation formula of the finite time state regulator. ResultsTh online identification of the ship motion model parameters and estimation of unknown disturbances are achieved, and a trajectory tracking control algorithm with "no need to worry after startup" is designed, reducing the strict requirements of parameter identification and control algorithms for experimental design. ConclusionUsing MATLAB to carry out numerical simulation and analyze the influence of weight matrix \begin{document}$ \boldsymbol{Q},\boldsymbol{R} $\end{document} and S on trajectory tracking accuracy, the results verify the effectiveness of the parameter identification and control algorithm.

With the vigorous development of the digital economy and the rapid construction of novel power system, deeply integrating advanced digital technology with power measurement technology and building a digital power measurement smart laboratory is an important way to accelerate the construction of a modern advanced measurement system and promote the digital transformation of power measurement. Focusing on the shortcomings and challenging problems in the construction of the current power measurement system, this paper describes the overall architecture and technical system of the power measurement smart laboratory in detail. On this basis, it conducts in-depth discussions on key issues such as measurement data panorama perception, edge computing and coordinated regulation, platform interaction and data processing, cross domain integration and value-added services, security protection, etc. during the construction of the laboratory, and looks forward to the realization of the functions and values of the digital power measurement smart laboratory, hoping to provide a certain reference for the development of power measurement technology.

In the rapidly evolving automotive electrification landscape, Belt-driven Starter and Generator (BSG) systems demand advanced motor modeling techniques. This research addresses critical challenges in six-phase hybrid excitation synchronous motors (6P-HESM) by developing a comprehensive framework that bridges theoretical modeling and practical automotive applications. A fusion modeling approach for 6P-HESM, leveraging finite-element analysis (FEA) and experimental data-driven is proposed in this paper to serve the application of 6P-HESM in BSG application. The proposed approach focuses on the fusion model construction, including analytical model derivation, model parameters acquisition with FEA and experimental data-driven and model fusion. The first one is carried out by vector space decomposition method with consideration of delta connection in stator windings. The second one is realized by experimental measurements, FEA and their combination with data-driven means, such as artificial neural networks. The last one is performed in MATLAB/Simulink platform based on the vehicle BSG scenario, and torque estimation is used as the target of model validation. Finally, the validation experiment is performed on test benches. The proposed model demonstrates superior torque estimation performance in BSG application without phase current. The proposed approach is also suitable for six-phase or three-phase permanent magnet synchronous motor, three-phase HESM and so on.

As one of the five major transportation systems, pipeline plays an important role in the energy transportation systems. In large-scale pipeline transportation systems, security issues such as leaks and explosions are prevalent, thus early detection of leaks is important to reduce security hazards in pipeline systems. Time series-based studies are widely used for leak detection in large-scale pipeline transportation systems, but single time-domain information, which ignores the spatial distribution of pressure sensors and does not consider periodic features, may not be sufficient for the detection accuracy of complex systems. To address it, a leak detection method based on frequency spatial-temporal convolution network (FSTCN) is proposed in this paper. Next, a spatial-encoder module for leak detection is proposed, which considers the spatial correlation of pressure sensors in pipeline systems. Second, a frequency-enhanced attention layer is proposed, which enables the feature extraction module to capture the periodic features of the pressure data. Meanwhile, a network self-updating mechanism is proposed which considers the changes in detection accuracy and data distribution to adapt to the continuously changing conditions of the pipeline systems. Finally, experiments are used to validate the proposed method, and nine time series classification models are chosen for comparison. The comprehensive results demonstrate that the effectiveness and superiority of the proposed leak detection method for large-scale pipeline systems.

Treatment based on syndrome differentiation is the core diagnostic and therapeutic thinking of traditional Chinese medicine (TCM), which is the key to determine clinical efficacy. Nowadays, research based on clinical data is the main method to explore the law of TCM syndrome and treatment, but the internal relationship of the key factors of "disease-syndrome-formula-medicine-effect" has not been truly and comprehensively analyzed, resulting in low clinical value of research results. Therefore, the author systematically sorted out the core problems of poor matching between electronic medical record and clinical research, the effect of data governance on data accuracy, difficulties to discover the law of TCM syndrome and treatment by data analysis methods. In addition, in the context of data driven, the big data platform of TCM clinical research should be established, and the data governance and analysis technology with artificial intelligence as the core should be developed, so as to realize the integration of clinical practice and research, providing new ideas and methods for the research of the law of TCM syndrome and treatment and promoting the development of TCM.

In recent years, with the rapid development of big data and artificial intelligence technology, data-driven automatic driving vehicle lane change intention recognition has become an active research area in the transportation field. Numerous studies have reported innovative and practical research results. However, this field still presents common technical challenges, such as accurately identifying the lane change process, handling missing lane change labels, and addressing imbalanced data categories. These issues remain the focal points of current research. This paper aims to classify and organize various data-driven methods, mainly focusing on lane change intention recognition methods based on traditional machine learning, deep learning, and ensemble learning. In the academic community, two primary approaches exist for identifying lane change behavior. The first approach mainly focuses on the vehicle not crossing the lane line, which is suitable for early recognition of the driver’s intention to change lanes. The second approach focuses on the actual crossing of lane markings by vehicles, which is often considered the complete lane change process. In academic research on lane change intention annotation, the selection of fixed time windows and heading angle thresholds plays a crucial role in the accuracy of annotation. These parameters affect the accurate recognition of lane change behavior and are directly related to the stability and reliability of autonomous driving and intelligent transportation system performance. Therefore, researchers have conducted in-depth investigations on the impact of these two parameters on annotation accuracy. To identify the optimal fixed time window and heading angle threshold, researchers have used the grid search optimization algorithm. This method performs well in fixed driving scenarios by traversing all possible parameter combinations and selecting the optimal parameters according to preset evaluation criteria. However, in practical applications, driving scenarios often exhibit diversity and complexity. Different driving environments, road conditions, and driving styles can impact the recognition of lane change intentions. Therefore, achieving adaptive parameter adjustment so that the annotation algorithm maintains high accuracy across various driving scenarios remains a challenging problem. To address the issue of imbalanced data categories in lane changing, researchers adopt two strategies. The first strategy involves adjusting the data sampling method, and under-sampling and oversampling techniques are used to balance the number of samples in each category. The second strategy involves the use of classification models with strong adaptability to imbalanced data, such as ensemble learning algorithms or cost-sensitive learning models, to maintain good classification performance.

Ferroelectric materials showing piezoelectricity, pyroelectricity and other functional properties have been found a variety of applications in electrical and electronic devices. These properties highly rely on polarization states in multi-scale structures, including ferroelectric domains mainly in mesoscopic scale, domain walls in microscopic scale and so on. However, it is still lack of effective method to characterize multi-scale polarization states simultaneously in ferroelectric materials. Here, we proposed a data-driven cross-scale polarization state recognition method based on scanning convergent beam electron diffraction (SCBED) to characterize the complicated polarization states in a PbZr 0.4 Ti 0.6 O 3 ceramic. This method employed a deep learning model to interpret the extensive dataset of CBED patterns generated during the scanning process and further validated by atomic resolution transmission electron microscope (ARTEM). The data-driven SCBED method provided a novel strategy for characterizing and interpreting the complicated cross-scale structure frame in ferroelectric materials. [ABSTRACT FROM AUTHOR]

The degree of rock mass discontinuity is crucial for evaluating surrounding rock quality, yet its accurate and rapid measurement at construction sites remains challenging. This study utilizes fractal dimension to characterize the geometric characteristics of rock mass discontinuity and develops a data-driven surrounding rock classification (SRC) model integrating machine learning algorithms. Initially, the box-counting method was introduced to calculate the fractal dimension of discontinuity from the excavation face image. Subsequently, crucial parameters affecting surrounding rock quality were analyzed and selected, including rock strength, the fractal dimension of discontinuity, the discontinuity condition, the in-situ stress condition, the groundwater condition, and excavation orientation. This study compiled a database containing 246 railway and highway tunnel cases based on these parameters. Then, four SRC models were constructed, integrating Bayesian optimization (BO) with support vector machine (SVM), random forest (RF), adaptive boosting (AdaBoost), and gradient boosting decision tree (GBDT) algorithms. Evaluation indicators, including 5-fold cross-validation, precision, recall, F1-score, micro-F1-score, macro-F1-score, accuracy, and the receiver operating characteristic curve, demonstrated the GBDT-BO model's superior robustness in learning and generalization compared to other models. Furthermore, four additional excavation face cases validated the intelligent SRC approach's practicality. Finally, the synthetic minority over-sampling technique was employed to balance the training set. Subsequent retraining and evaluation confirmed that the imbalanced dataset does not adversely affect SRC model performance. The proposed GBDT-BO model shows promise for predicting surrounding rock quality and guiding dynamic tunnel excavation and support. [ABSTRACT FROM AUTHOR]

The manufacturing industry must maintain high-quality standards while meeting customer demands for customization, reduced carbon footprint, and competitive pricing. To address these challenges, companies are constantly improving their production processes using quality management tools. A crucial aspect of this improvement is the root cause analysis of manufacturing defects. In recent years, there has been a shift from traditional knowledge-driven approaches to data-driven approaches. However, there is a gap in the literature regarding a systematic overview of both methodological types, their overlaps, and the challenges they pose. To fill this gap, this study conducts a scoping literature review of root cause analysis in manufacturing, focusing on both data-driven and knowledge-driven approaches. For this, articles from IEEE Xplore, Scopus, and Web of Science are examined. This review finds that data-driven approaches have become dominant in recent years, with explainable artificial intelligence emerging as a particularly strong approach. Additionally, hybrid variants of root cause analysis, which combine expert knowledge and data-driven approaches, are also prevalent, leveraging the strengths of both worlds. Major challenges identified include dependence on expert knowledge, data availability, and management issues, as well as methodological difficulties. This article also evaluates the potential of artificial intelligence and hybrid approaches for the future, highlighting their promises in advancing root cause analysis in manufacturing. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Graphics is the property of Journal of Graphics Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Computer Measurement & Control is the property of Magazine Agency of Computer Measurement & Control and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Railway Science & Engineering is the property of Journal of Railway Science & Engineering Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Background: The blood supply chain (BSC) is crucial for providing safe and sufficient blood, but it faces numerous challenges and needs to be robust and resilient. This study provides a comprehensive model for managing and optimizing the BSC in real-world scenarios, including emergency and routine circumstances and with consideration of health equity concepts. Method: Classic time-series models are applied to predict future supply chain circumstances, addressing uncertainty in blood demand and the need for timely supply. A structured framework and medical preferences are prioritized to optimize distribution, minimize blood shortages, minimize wastage due to expiry, and maximize blood freshness. Genetic algorithms (GA) and particle swarm optimization (PSO) are used to solve mathematical models quickly and efficiently, ensuring reliable operation. Result: The model's outcomes can effectively meet the daily needs of the BSC and assist decision-makers managing blood inventory and distribution, improving robustness and resilience. Conclusions: Utilizing weights allows for the effective management of each objective function to convert the model into a single-objective mixed-integer linear programming (SO-MILP) based on unique conditions, enabling the system to self-adjust for optimal performance, boosting the sustainability of the blood supply chain, and promoting the principle of health equity under diverse real-world settings. [ABSTRACT FROM AUTHOR]