Combining Machine Learning Research Articles

Multiscale groundwater level forecasts with multi-model ensemble approaches: Combining machine learning models using decision theories and bayesian model averaging

Creating precise groundwater level (GWL) prediction models is of crucial significance for the productive use, extended planning, and controlling of limited sub-surface water supplies. In this research, the accuracy of GWL forecasts in Bangladesh was enhanced for three weeks by utilizing ensembles of Machine Learning (ML) models. Six advanced ML-based models were developed and assessed using eight performance indices, and an Overall Ranking (OR) was provided by combining the rankings produced by Grey Relational Analysis (GRA), Variation Coefficient (COV), and Shannon's Entropy (SE). The standalone forecasting models demonstrated excellent performance across the three forecasting horizons, with accuracy values ranging from 0.986 to 0.997 for one-step, 0.971 to 0.999 for two-step, and 0.960 to 0.997 for three-step forecasts at GT3330001. Results also revealed that three ranking techniques (SE, COV, and GRA), as well as their combined ranking (OR), produced different best-performing models at different prediction horizons for different observation wells. Weighted average ensembles of the prediction models were developed by calculating individual model weights using four ensemble modelling techniques: SE, COV, GRA, and Bayesian Model Averaging (BMA). The BMA-based ensemble technique outperformed three benchmark ensemble approaches, achieving R = 0.947, KGE = 0.925, IOA = 0.972, MAE = 0.062 m, and RMSE = 0.123 m for one-step-ahead forecasts at GT3330001. The findings exhibit a consistent trend across other forecasting horizons and observation wells. Finally, the Dempster-Shafer evidence theory was employed to rank the single and composite models. The ranking results demonstrated that the BMA-based ensemble consistently secured the top position (with the weight values of 0.997, 0.991, and 0.987 for one-week, two-weeks, and three-weeks forward forecasts at GT3330001) for all forecasting horizons and observation wells. This study shows that the BMA-based composite model can produce more accurate GWL projections at Bangladesh study location, with potential for application in other regions worldwide.

CMShark: A NetFlow and machine-learning based crypto-jacking intrusion-detection method

Crypto-jacking attack is a novel type of cyber-attack on the internet that has emerged because of the popularity of digital currencies. These attacks are the most common type of attacks in the cryptocurrency field because of their specific features such as easy scenario, un-traceability, and ease of secrecy. In crypto-jacking attacks, it is common to embed malicious code inside website scripts. Different techniques have been provided to deal with Crypto-jacking attacks, but crypto-jacking attackers bypass them by limiting resources. The crypto-mining services provided on the internet are legal, and due to the anonymous nature of cryptocurrencies, client identification is a challenging task. Improving the accuracy and performance of the Crypto-jacking attack detection methods are the main objectives of this study. In this paper, a hybrid network-based method to identify these attacks to achieve better and more accurate results. The proposed solution (CMShark) is a combination of machine learning (ML) models, IP blacklisting and payload inspection methods. In the ML model, the packets are classified using size patterns; in IP blacklisting, attacks are detected based on known infected addresses and infected scripts. In payload inspection, the provided information on the packet payload is searched for any suspicious keywords. The proposed method relies solely on the network and is deployed on the edge of the network, making it infrastructureindependent. The proposed detection model reaches an accuracy score of 97.02%, an F1-score of 96.90% a ROC AUC score of 97.20% in input NetFlow classification; and a 93.98% accuracy score, 94.30% F1-score and 97.30% ROC AUC score in output NetFlow classification.

PSX-22 Boosting techniques for prediction of dry matter intake of cattle in confinement settings and SHAP analysis on the models’ output

Abstract Measurement of individual dry matter intake (DMI) is critical to improving the management of the beef herd and is necessary to determine the feed efficiency of individual cattle. Combining machine learning models and proxies for DMI (e.g., water intake, body weight, climatic variables) has the potential to replace the need for dedicated feed intake equipment. It would also open the possibility of measuring DMI in situations like grazing where quantifying DMI directly is not possible. In this study, we examined the relationship between DMI, animal variables and environmental variables. In this study, we developed and tested machine learning models based boosting techniques to develop approaches to predict DMI in cattle, using the same data we have previously published results from Random Forest techniques in the past. The dataset included both bulls and steers. The developed models include both regular boosting regression models and random effects boosting models. The regular boosting technique models such as gradient boosting regression, light gradient boosting regression, extreme gradient boosting regression. Similarly, random effects boosting machine learning model Gaussian process boosting (GPBoost) was developed to consider the random effects associated with the intake behavior of the animals. The extreme gradient boosting model outperformed other regular boosting regression models with a r2 of 0.81 and 0.48 on training and testing datasets respectively. However, it does not consider the random effects associated with the longitudinal nature of the animal intake and weather variables and prone to overfitting. GPBoost model variants were tested and developed to avoid overfitting and to incorporate a random effects model, which we propose as the overall best performing model. The GPBoost generalized well on the training and testing datasets with r2 scores of 0.55 and 0.54 respectively. To further understand the model’s performance on the unseen dataset, we used SHAP analysis on the GPBoost. The SHAP analysis revealed that water intake, full body weight and age of the animals contributed significantly to predicting unseen data.

State estimation of a biogas plant based on spectral analysis using a combination of machine learning and metaheuristic algorithms

The continuous monitoring of the state variables of a biogas plant remains a challenge due to the necessity of an appropriate measuring device. The collection, transportation, and laboratory measurement of the biogas sample are required, and regularly performing these activities is expensive and time-consuming. The objective of this study is to investigate a potential solution for the real-time monitoring of the state variables of a biogas plant, which involves the integration of spectral data from near-infrared (NIR) sensors with advanced machine learning (ML) and metaheuristic algorithms. A total of 635 samples were prepared in a laboratory and subsequently analyzed using portable NIR sensors mounted in sensor brackets that were 3D-printed for this study. The resulting spectral data were subjected to several preprocessing processes, and this study combines numerous methods rather than selecting the optimal one. Several ML models were then trained on the processed data, which have a realistic range as in the actual biogas plant. The developed algorithm exhibits an accuracy level greater than 82 % when classifying the volatile fatty acids (VFA)/total alkalinity (TA) ratio and acetic acid concentration of biogas samples. Furthermore, the dry matter content of biogas samples can be accurately predicted with an RMSE of approximately 1.6 %. In addition, the significance of selecting suitable hyperparameters is demonstrated, which may substantially influence the outcome. These findings indicate that NIR sensors are a realistic option for monitoring the operational status of biogas plants.

Cardiovascular disease diagnosis: a holistic approach using the integration of machine learning and deep learning models

BackgroundThe incidence and mortality rates of cardiovascular disease worldwide are a major concern in the healthcare industry. Precise prediction of cardiovascular disease is essential, and the use of machine learning and deep learning can aid in decision-making and enhance predictive abilities.ObjectiveThe goal of this paper is to introduce a model for precise cardiovascular disease prediction by combining machine learning and deep learning.MethodTwo public heart disease classification datasets with 70,000 and 1190 records besides a locally collected dataset with 600 records were used in our experiments. Then, a model which makes use of both machine learning and deep learning was proposed in this paper. The proposed model employed CNN and LSTM, as the representatives of deep learning models, besides KNN and XGB, as the representatives of machine learning models. As each classifier defined the output classes, majority voting was then used as an ensemble learner to predict the final output class.ResultThe proposed model obtained the highest classification performance based on all evaluation metrics on all datasets, demonstrating its suitability and reliability in forecasting the probability of cardiovascular disease.

Environmental pollutant Di-(2-ethylhexyl) phthalate induces asthenozoospermia: new insights from network toxicology.

The global decline in sperm quality in men is closely associated with environmental exposure to the plasticizer Di-(2-ethylhexyl) phthalate (DEHP), but the molecular mechanisms underlying its induction of asthenozoospermia (AZS) remain incompletely understood. By integrating the toxicological targets of DEHP and differential genes in AZS patients, and combining machine learning, molecular docking, and dynamics simulations, this study successfully identified hub genes and signaling pathways induced by DEHP in AZS, aiming to provide new strategies for the prevention and treatment of this disease. A total of 26 toxicological targets were identified, with FGFR1, MMP7, and ST14 clearly defined as playing crucial regulatory roles in DEHP-induced AZS. This study also reveals that DEHP may induce reproductive system inflammation, affecting the proliferation and survival of reproductive cells, and subsequently impacting sperm vitality, possibly through regulating the mTORC1 pathway, TNF-α signaling via the NF-κB pathway, and MYC targets v1 pathway. Furthermore, changes in the immune microenvironment revealed the significant impact of immune status on testicular function. In conclusion, this study provides important scientific evidence for understanding the molecular mechanisms of AZS and developing prevention and treatment strategies based on toxicological targets.

Pharmaceutical aerosol transport in airways: A combined machine learning (ML) and discrete element model (DEM) approach

The continuum and discrete phase interaction could significantly affect the inhaled particle's transport behaviour. The accurate analysis of continuum and discrete phase interaction needs computational fluid dynamics (CFD) and Discrete Element Method (DEM) simulation, which is computationally expensive. Therefore, this study aims to develop a novel machine learning (ML) prediction model from CFD-DEM data to predict pharmaceutical aerosol transport in airways accurately. This study uses the CFD model for the continuum and DEM for the discrete phases. A soft sphere approach was used to calculate the overlap of the colliding particles. Proper validation was performed to ensure the accuracy of the present model. The CFD-DEM model analysed the particle transport in an idealised and realistic airway model, and different methods were used to analyse the transport behaviour. As the flow rate increased from 15 to 60 lpm, the deposition efficiency (DE) significantly improved due to particle interaction, rising from 20 % to 42 % without interaction and from 50 % to a remarkable 76 % with interaction, demonstrating the critical role of particle interaction in enhancing DE across varying flow rates. During the particle-particle interaction, a stagnation point and a high-pressure zone were observed at the airway model's carinal angle. Finally, a ML prediction model is developed from CFD-DEM data, which accurately predicts the pharmaceutical aerosol deposition in airways. The ML prediction model predicts the deposition pattern for different flow rates and particle sizes without CFD-DEM simulations. The present findings and more case-specific investigation would advance the knowledge of aerosol transport in airways and benefit more efficient targeted drug delivery devices.

Understanding Learning from EEG Data: Combining Machine Learning and Feature Engineering Based on Hidden Markov Models and Mixed Models.

Theta oscillations, ranging from 4-8 Hz, play a significant role in spatial learning and memory functions during navigation tasks. Frontal theta oscillations are thought to play an important role in spatial navigation and memory. Electroencephalography (EEG) datasets are very complex, making any changes in the neural signal related to behaviour difficult to interpret. However, multiple analytical methods are available to examine complex data structures, especially machine learning-based techniques. These methods have shown high classification performance, and their combination with feature engineering enhances their capability. This paper proposes using hidden Markov and linear mixed effects models to extract features from EEG data. Based on the engineered features obtained from frontal theta EEG data during a spatial navigation task in two key trials (first, last) and between two conditions (learner and non-learner), we analysed the performance of six machine learning methods on classifying learner and non-learner participants. We also analysed how different standardisation methods used to pre-process the EEG data contribute to classification performance. We compared the classification performance of each trial with data gathered from the same subjects, including solely coordinate-based features, such as idle time and average speed. We found that more machine learning methods perform better classification using coordinate-based data. However, only deep neural networks achieved an area under the ROC curve higher than 80% using the theta EEG data alone. Our findings suggest that standardising the theta EEG data and using deep neural networks enhances the classification of learner and non-learner subjects in a spatial learning task.

Remote Patient Monitoring and Machine Learning in Acute Exacerbations of Chronic Obstructive Pulmonary Disease: Dual Systematic Literature Review and Narrative Synthesis.

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are associated with high mortality, morbidity, and poor quality of life and constitute a substantial burden to patients and health care systems. New approaches to prevent or reduce the severity of AECOPD are urgently needed. Internationally, this has prompted increased interest in the potential of remote patient monitoring (RPM) and digital medicine. RPM refers to the direct transmission of patient-reported outcomes, physiological, and functional data, including heart rate, weight, blood pressure, oxygen saturation, physical activity, and lung function (spirometry), directly to health care professionals through automation, web-based data entry, or phone-based data entry. Machine learning has the potential to enhance RPM in chronic obstructive pulmonary disease by increasing the accuracy and precision of AECOPD prediction systems. This study aimed to conduct a dual systematic review. The first review focuses on randomized controlled trials where RPM was used as an intervention to treat or improve AECOPD. The second review examines studies that combined machine learning with RPM to predict AECOPD. We review the evidence and concepts behind RPM and machine learning and discuss the strengths, limitations, and clinical use of available systems. We have generated a list of recommendations needed to deliver patient and health care system benefits. A comprehensive search strategy, encompassing the Scopus and Web of Science databases, was used to identify relevant studies. A total of 2 independent reviewers (HMGG and CM) conducted study selection, data extraction, and quality assessment, with discrepancies resolved through consensus. Data synthesis involved evidence assessment using a Critical Appraisal Skills Programme checklist and a narrative synthesis. Reporting followed PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. These narrative syntheses suggest that 57% (16/28) of the randomized controlled trials for RPM interventions fail to achieve the required level of evidence for better outcomes in AECOPD. However, the integration of machine learning into RPM demonstrates promise for increasing the predictive accuracy of AECOPD and, therefore, early intervention. This review suggests a transition toward the integration of machine learning into RPM for predicting AECOPD. We discuss particular RPM indices that have the potential to improve AECOPD prediction and highlight research gaps concerning patient factors and the maintained adoption of RPM. Furthermore, we emphasize the importance of a more comprehensive examination of patient and health care burdens associated with RPM, along with the development of practical solutions.

Combining machine learning (ML) and participatory rural appraisal (PRA) for disaster risk preparedness (DRP): Evidence from the poorest region of Luzon, Philippines

In the field of social science, disaster risk preparedness (DRP) is considered immeasurable due to its multidimensional nature, making it infamously difficult to quantify. The current measurements are costly, labor-intensive, and time-consuming. Consequently, policymakers struggle to target policies effectively when implementing disaster risk reduction management initiatives. By combining Participatory Rural Appraisal (PRA) and Machine Learning (ML) to train and test community-based system datasets, this work proposes novel approaches to DRP in the poorest region of Luzon, Philippines. We utilized sophisticated econometrics models along with ML categorization methods. Through the analysis of 34 locales and 4 sectors within a disaggregation system over 429 ensemble runs using cross-validation techniques, we then combined the results. The Support Vector Machine (SVM) classifier achieved the highest accuracy of 91.55 % randomly and 94.53 % within the pipeline, surpassing all other models. It also confirms the current relationship between DRP and multidimensional attributes (a total of 21 factors) in terms of correlation and causation. Our work showcases the potential of ML for disaster risk prediction, potentially reducing costs, saving labor, and optimizing time, especially in the most impoverished areas of the Philippines. Ultimately, through extensive PRA, the outcomes have provided different localities with tools for targeting policies in disaster risk management.

Scalable and Robust Fraud Detection in Distributed Systems

The rise of distributed systems has increased the need for advanced fraud detection mechanisms. Cybercriminals increasingly exploit the distributed and decentralized nature of these systems, posing challenges for traditional fraud detection techniques that rely on centralized data analysis. In this paper, we propose a novel approach to fraud detection that is decentralized, scalable, and capable of real-time detection across diverse nodes in distributed systems. Our solution combines machine learning techniques, including anomaly detection and classification algorithms, with decentralized consensus mechanisms. We evaluate our approach using a large-scale financial dataset and outline its performance in terms of accuracy, latency, and scalability. This work also discusses challenges such as data privacy, adversarial attacks, and regulatory compliance, providing directions for future research

Toxicity prediction and risk assessment of per- and polyfluoroalkyl substances for threatened and endangered fishes

Per- and polyfluoroalkyl substances (PFASs) are severely polluted in aquatic environments and can harm aquatic organisms. Due to the limitation of conducting toxicity experiments directly on threatened and endangered (T&E) species, their toxicity data is scarce, hindering accurate risk assessments. The development of computational toxicology makes it possible to assess the risk of pollutants to T&E fishes. This study innovatively combined machine learning models, including random forest (RF), artificial neural network (ANN), and XGBoost, and the QSAR-ICE model to predict chronic developmental toxicity data of PFASs to T&E fishes. Among these, the XGBoost model exhibited superior performance, with R2 of 0.95 and 0.81 for the training and testing sets, respectively. Internal and external validation further confirmed that the XGBoost model is robust and reliable. Subsequently, it was used to predict chronic developmental toxicity data for seven priority PFASs to T&E fishes in the Yangtze River. Acipenseridae fishes (e.g., Acipenser dabryanus and Acipenser sinensis) showed high sensitivity to PFASs, possibly due to their unique lifestyle and physiological characteristics. Based on these data, the predicted no-effect concentration (PNEC) of individual PFASs was calculated, and the risk for T&E fishes in the Yangtze River was assessed. The results indicated that the risk of PFASs to T&E fishes is low (3.85 × 10−9∼8.20 × 10−4), with perfluorohexanoic acid (PFHxA) and perfluorooctanoic acid (PFOA) as the high-risk pollutants. The risk in the middle and lower reaches of the river is higher than in the upper reaches. This study provides a new approach for obtaining chronic toxicity data and conducting risk assessments for T&E species, advancing the protection of T&E species worldwide.

Turbidity assessment in coastal regions combining machine learning, numerical modeling, and remote sensing

ABSTRACT Machine learning models for water quality prediction often face challenges due to insufficient data and uneven spatial-temporal distributions. To address these issues, we introduce a framework combining machine learning, numerical modeling, and remote sensing imagery to predict coastal water turbidity, a key water quality proxy. This approach was tested in the Great Lakes region, specifically Cleveland Harbor, Lake Erie. We trained models using observed and synthetic data from 3D numerical models and tested them against in situ and remote sensing data from PlanetLabs' Dove satellites. High-resolution (HR) data improved prediction accuracy, with RMSE values of 0.154 and 0.146 log10(FNU) and R2 values of 0.92 and 0.93 for validation and test datasets, respectively. Our study highlights the importance of unified turbidity measures for data comparability. The machine learning model demonstrated skill in predicting turbidity through transfer learning, indicating applicability in diverse, data-scarce regions. This approach can enhance decision support systems for coastal environments by providing accurate, timely predictions of water quality variables. Our methodology offers robust strategies for turbidity and water quality monitoring and holds significant potential for improving input data quality for numerical models and developing predictive models from remote sensing data.

Combining Machine Learning, Embedded Sensor Networks and Additive Burner Design for Combustor Structural Health Monitoring

Abstract The crystAIr project is about sensitising burners for flame monitoring based on a discrete number of dynamic pressure sensors and a machine learning approach. The context is the development of better aircraft engines and the prospect of technical solutions for hydrogen-fuelled gas turbines. The combination of sensors and appropriate signal processing is designed to mimic a ‘feel’ for the state of the flame. The idea is to monitor the gas turbine's correct operation in real-time and anticipate impending problems such as flame blowout, combustion instability or flashback. Compared to conventional flame monitoring based on signal sampling, thresholding, band-pass analysis and time-lag measurement, this method should be less computationally intensive, more sensitive and more robust to artefacts. Not only should it be more responsive, but it should also anticipate problems based on a lessons-learned approach and outperform the conventional precursors. An unassisted machine learning method is used to achieve this. A custom-built AM burner designed for this experiment provides multiple instrumentation options. A powerful data acquisition system is set up to collect data on multiple channels, over a long time duration and at high speed to collect the learning signal chunks. The focus is on the detection of flames and, more specifically, the moment of their ignition and extinction, the operating conditions that are prone to flashback and the stability of the combustion. Additional measurement techniques are used to refine the learning method. The paper covers all these points and presents the first results.

Transplanting network pharmacology technology into food science research: A comprehensive review on uncovering food-sourced functional factors and their health benefits.

Network pharmacology is an emerging interdisciplinary research method. The application of network pharmacology to reveal the nutritional effects and mechanisms of active ingredients in food is of great significance in promoting the development of functional food, facilitating personalized nutrition, and exploring the mechanisms of food health effects. This article systematically reviews the application of network pharmacology in the field of food science using a literature review method. The application progress of network pharmacology in food science is discussed, and the mechanisms of functional factors in food on the basis of network pharmacology are explored. Additionally, the limitations and challenges of network pharmacology are discussed, and future directions and application prospects are proposed. Network pharmacology serves as an important tool to reveal the mechanisms of action and health benefits of functional factors in food. It helps to conduct in-depth research on the biological activities of individual ingredients, composite foods, and compounds in food, and assessment of the potential health effects of food components. Moreover, it can help to control and enhance their functionality through relevant information during the production and processing of samples to guarantee food safety. The application of network pharmacology in exploring the mechanisms of functional factors in food is further analyzed and summarized. Combining machine learning, artificial intelligence, clinical experiments, and in vitro validation, the achievement transformation of functional factor in food driven by network pharmacology is of great significance for the future development of network pharmacology research.

Predictors of High Healthcare Cost Among Patients with Generalized Myasthenia Gravis: A Combined Machine Learning and Regression Approach from a US Payer Perspective.

High healthcare costs could arise from unmet needs. This study used random forest (RF) and regression methods to identify predictors of high costs from a US payer perspective in patients newly diagnosed with generalized myasthenia gravis (gMG). Adults with gMG (first diagnosis = index) were selected from the IQVIA PharMetrics® Plus database (2017-2021). Predictors of high healthcare costs were measured 12 months pre-index (main cohort) and during both the 12 months pre- and post-index (subgroup). Top 50 predictors of high costs [≥ $9404 (main cohort) and ≥$9159 (subgroup) per-patient-per-month] were identified with RF models; the magnitude and direction of association were estimated with multivariable modified Poisson regression models. The main cohort and subgroup included 2739 and 1638 patients, respectively. In RF analysis, the most important predictors of high costs before/on the index date were index MG exacerbation, all-cause inpatient admission, and number of days with corticosteroids. After the index date, these were immunoglobulin and monoclonal antibody use and number of all-cause outpatient visits and MG-related encounters. Adjusting for the top 50 predictors, post-index immunoglobulin use increased the risk of high costs by 261%, monoclonal antibody use by 135%, index MG exacerbation by 78%, and pre-index all-cause inpatient admission by 27% (all p < 0.05). This analysis links patient characteristics both before the formal MG diagnosis and in the first year to high future healthcare costs. Findings may help inform payers on cost-saving strategies, and providers can potentially shift to targeted treatment approaches to reduce the clinical and economic burden of gMG.

Efficient grid management: smart forecasting of short-term power load using PSO-LSTM

Recent load forecasting techniques combining machine learning models and hyperparameter optimization algorithms have shown success for short-term load forecasting (STLF) task, but they often require complex programming, higher computational costs, and greater parameter tuning. In this paper, we introduce an improved STLF model that combines Long Short-Term Memory (LSTM) neural network with Particle Swarm Optimization (PSO) for enhanced performance. In the proposed approach, the number of hidden neurons in different LSTM layers, learning rate and the number of iterations for training are optimized using the PSO algorithm. To validate the effectiveness of this method, meteorological data and historical load data from a real-world power grid are used as input. The experimental results reveal PSO significantly enhances hyperparameter tuning for LSTM neural networks, leading to improved predictive modelling. The PSO-LSTM model performed better than the LSTM model by more than 20% (in terms of Mean Absolute Error), and showed low sensitivity to hyperparameters. Comparative analysis with alternative approaches from the literature further validates the PSO-LSTM’s effectiveness in STLF. Additionally, the model achieved stable multi-step prediction capabilities, with average errors of 3.6445 for MAE, 4.6509 for RMSE, and 4.6519 for MAPE over a 1–4 day ahead lead times. This study highlights PSO-LSTM’s enhanced robustness and accuracy in power load prediction while addressing hyperparameter tuning challenges through self-optimization.

Sealing rubber ring design based on machine learning algorithm combined progressive optimization method

In this study, a progressive optimization method combining machine learning and optimization method is proposed and applied to seal structure design. The method is divided into two stages to select the optimal design gradually. So as to find the best design scheme meeting the design requirements. For optimal design, numerical calculation method is commonly used, but hard to evaluate the optimal solution. In this work, a series of numerical model considering the effect of super elastic material about O-ring study the waterproof performance behavior of a rubber seal. K nearest neighbors (KNN) of machine learning algorithms applied to the simulation data to predict the appropriate bolt pretension classification. Furthermore, use TOPSIS method to optimize the groove depth of 30 N bolt pretension classification. By using the TOPSIS method to consider the stress of the rubber component, optimization analysis is conducted to find the optimal design. Results show that the dual optimization method can quickly predict the best design scheme. Through the experiment, a prototype test under the condition of IPX7 verify the method. The design scheme selected by this method meets the waterproof grade requirements. There are no water stains on the surface of the O-ring and inside the motor. This paper provides a fast optimization design method for the design of sealing structure.

Electroreduction of nitrate to ammonia on graphyne-based single-atom catalysts by combined density functional theory and machine learning study

Nitrate pollution has emerged as a significant global issue, prompting increased interests in electrocatalytic reduction of nitrate to ammonia (NO3RR). This study explores the use of graphyne as the substrate to anchor single transition metal (TM) atoms from the 3d and 4d series on acetylene corners, creating single-atom catalysts (SAC) for the NO3RR reaction. Two catalysts, CrGY and MoGY, were finally identified as promising catalysts for NO3RR with low limiting potential and high selectivity for NH3. To further investigate the origin of catalyst activity, in addition to traditional electronic structrue analysis, machine learning models of the least solution shrinkage and selection operator (LASSO), the sure independence screening and sparsifying operator (SISSO) and the gradient boosting regression (GBR) algorithm are adopt to reveal the intrinsic properties of catalysts that determine the NO3RR performance. After preliminary evaluation of the 13 prepared features with LASSO algorithm, GBR models revealed that the nitrate adsorption free energy ΔGNO3- and the NO bond length are the two most important descriptors for NO3RR performance. A general equation for the relationship between the limiting potential and the characteristics of the catalysts are obtained by SISSO model. Finally, with the assitance of combined density functional calculations and machine learning, this study provides new enlightening insights into practical application of graphyne as SACs for NO3RR.

Algorithms and strategies for fraud prevention on online platforms

The topic of fraud prevention on online platforms is an urgent area in the field of information security and user protection. Modern fraud prevention algorithms and strategies include an integrated approach combining machine learning, analytical tools, and multi-layered protection systems. The key aspects are the detection of anomalies in user behavior, the use of machine learning algorithms to identify suspicious activities, as well as the integration of data verification and verification systems. Behavioral analytical models that help predict and prevent potential threats also play an important role in the fight against fraud. The effectiveness of these methods and strategies depends on their ability to adapt to evolving threats and ensure a balance between security and user convenience.