Model Predictions Research Articles

Machine Learning-Guided Selection of Cyclodextrins for Enhanced Biosynthesis and Capture of Volatile Terpenes.

Nootkatone and limonene are valuable volatile organic compounds (VOCs), but their biosynthetic production is hindered by volatility. This study employed machine learning to guide cyclodextrin (CD) selection for encapsulating these VOCs, with a focus on nootkatone capture during fermentation to prevent losses and potentially replace dodecane as an organic solvent extractant. A LightGBM model accurately predicted complexation free energies (ΔG) between CDs and guest molecules (R2 = 0.80 on a 10% test set, with a mean absolute error of 1.31 kJ/mol and a root-mean-squared error of 1.90 kJ/mol). Experimental ranking of 7 CD types validated the model's ΔG predictions and encapsulation performance rankings. Nootkatone showed high encapsulation efficiencies ranging from 21.29% (α-CD) to 88.41% (Me-β-CD), capturing 22.61-116.71 mg/g CD. Notably, Hp-γ-CD, which is the least studied or used CD in research, performed well with nootkatone (63.64%, 84.01 mg/g CD) despite model discrepancies. For limonene, encapsulation efficiencies spanned from 0.62% (Hp-γ-CD) to 55.45% (β-CD), with 0.61-84.28 mg/g CD encapsulated. Constructed engineered Saccharomyces cerevisiae strains produced nootkatone (up to 97.30 mg/L captured by 10 mM Me-β-CD) from de novo fermentation using glucose as a carbon source. This approach demonstrated the potential of CDs to replace dodecane as an organic solvent for terpene extraction during fermentation. The study highlights machine learning's potential for guiding CD selection to enhance volatile terpene biosynthesis, capture, and utilization during fermentation, offering a more environmentally friendly alternative to traditional organic solvent-based extraction methods.

Improving predictability of post-operative knee joint line obliquity in high tibial osteotomy: A new radiographic parameter correcting for foot position.

Joint line obliquity is crucial for the long-term success of a high tibial osteotomy (HTO), particularly when large corrections are made. The relationship between the size of correction and changes in joint line obliquity (KJLO) is complex, often leading to the preference for a double-level osteotomy to manage significant post-operative obliquity. We aim to improve the prediction of knee joint line orientation changes and correction size by examining the influence of foot position. The goal is to develop a linear model to predict post-operative KJLO, which could assist in determining whether a single- or double-level osteotomy is necessary to prevent excessive post-operative joint line obliquity. A retrospective radiographic analysis was conducted on patients who underwent HTO surgery between April 2016 and April 2017. Ninety-one patients were randomly selected for radiographic measurements on long-leg radiographs pre- and 3-month post-operative. A novel radiographic parameter was introduced to realign the foot of the operated side to the midline. The foot position relative to the midline was assessed by determining the angle formed when rotating the natural foot position onto the midline using the centre of the hip as a rotation point. The angle obtained by subtracting this from the KJLO results in a corrected KJLO for midline foot position (aKJLO). Predictive model of differences between pre- and post-operative values of MPTA (ΔMPTA [medial proximal tibial angle]) and KJLO (ΔKJLO) was less predictive (0.185 [p < 0.001]) than ΔMPTA and ΔKJLO corrected for foot position, namely ΔaKJLO (0.688 [p < 0.001]). Adding more parameters did not significantly improve the linear model's predictions. Predictability of aKJLO could be significantly enhanced. With this new parameter, inter- and intra-variability of foot position is bypassed. A safe MPTA position of 91° can be presumed; however, inter-individual variability in limb adaptation following correction remains uncertain. Level IV.

Safer Floors in Public Service Buildings Based on Machine Learning

Abstract This study investigates the effects of different floor surfaces on slip safety in public service buildings (PSBs) with heavy pedestrian traffic. The K-means clustering method is used to classify various floor types and slip safety risks. The dynamic friction coefficient (DCOF) for floor coverings, such as natural stone, ceramic, laminate, and PVC, were measured in both dry and wet conditions across thirty public institutions. These measurements were obtained using the GMG 200 and Wessex S885 pendulum testers, providing a comprehensive assessment of the slip resistance of these surfaces. The machine learning models employed in the study were XGBoost, K-Nearest Neighbors (KNN), and SVC. The models were evaluated using 5-fold cross-validation. The analysis revealed that the most significant parameter in DCOF predictions for the XGBoost model was environmental conditions (EC). Performance analysis showed that the SVC model achieved the highest F1 score (0.75 ± 0.01) and AUC value (0.83), outperforming the other models. Additionally, DCOF values from slip tests were grouped into five clusters using the K-means method, and a slip safety risk scale was developed. Statistically significant differences were observed in DCOF values based on usage areas, environmental conditions, test methods, and surface materials. For instance, hospital floors were found to be generally safe in dry conditions but posed a risk in wet conditions. Based on these findings, actionable safety measures were suggested, such as applying anti-slip coatings in high-risk areas, selecting flooring materials with higher DCOF values for moisture-prone environments, and implementing regular slip resistance testing to maintain safety standards. In conclusion, this study demonstrates that machine learning models can effectively assess the slip resistance of floor surfaces. The findings offer valuable guidance for construction industry professionals and researchers in improving safety measures and minimizing slip risks. Future research with larger datasets and diverse conditions could enhance the understanding of this issue and further improve model performance.

Constraints on neutrino nonstandard interactions from COHERENT, PandaX-4T and XENONnT

We investigate constraints on neutrino nonstandard interactions (NSIs) in the effective field theory framework, using data from the first measurement of solar B8 neutrinos via coherent elastic neutrino-nucleus scattering (CEνNS) in the PandaX-4T and XENONnT experiments and data from the COHERENT experiment. The impacts of neutrino NSIs on the CEνNS cross section and the matter effect in the propagation of solar neutrinos are included, while we obtain that the expected number of CEνNS events is more sensitive to neutrino NSIs appearing in the cross section. Due to relatively large statistical uncertainties, the sensitivities of the PandaX-4T and XENONnT experiments to the neutrino NSIs are currently limited, compared to the COHERENT experiment. Besides, we find that since the central value of the measured CEνNS counts significantly differs from the Standard Model prediction, the sensitivity of PandaX-4T experiment is even more restricted compared to XENONnT. However, the measurements of PandaX-4T and XENONnT are uniquely sensitive to the neutrino NSIs for the τ flavor due to oscillation feature of the solar B8 neutrinos. We also assess how the experimental central value, exposure, and systematic uncertainties will affect the constraints on neutrino NSIs from various CEνNS measurements in the future. Published by the American Physical Society 2025

Graded Walking Energetics under Cold Strain.

Although research supports selecting steeper routes to minimize metabolic rate (Ṁ) in mountain racers, the "steeper is cheaper" strategy has yet to be confirmed for slower, more typical graded walking speeds under cold stress. Confirm whether "steeper is cheaper" is true for typical graded walking speeds in individuals exposed to incremental cold stress. Fourteen healthy, military-age adults (age, 24 ± 6 years; height, 1.72 ± 0.08 m; body mass, 72 ± 16 kg) completed four 20-min treadmill walks in three ambient temperatures (20, 10, and 0 °C) in light clothing (i.e., shorts, t-shirt, light gloves). Each walk involved five stages at incremental vertical speeds (0.00, 1.93, 3.86, 5.79, 7.79 m·min-1) but variable treadmill speeds (0.54, 0.72, 1.07 m·s-1). To verify the "steeper is cheaper" strategy, we compared Ṁ between treadmill speeds at matched vertical speeds in each temperature. We also tested if the 90% confidence interval around the mean percent paired difference between Load Carriage Decision Aid (LCDA) metabolic model predictions and measured Ṁ was within ±10%. Ṁ was significantly higher for the faster treadmill speed at matched vertical speeds in all but two comparisons at 20 °C, all but two comparisons at 10 °C, and all comparisons at 0 °C (p < 0.05). LCDA metabolic model predictions were statistically equivalent to measured Ṁ during graded walking at 20 °C (90% CI, -3.1, 0.6%) and 10 °C (-7.8, -2.6%) but not 0 °C (-16.2, -9.5%). Route planners should recommend steeper but shorter routes to minimize Ṁ in individuals that walk in temperate-to-cold environments. The LCDA metabolic model provides accurate Ṁ predictions in lightly dressed individuals in temperatures down to 10 °C, but users should expect underestimated Ṁ in 0 °C or colder.

Context-Based Human Influence and Causal Responsibility for Assisted Decision-Making.

The impact of the context in which automation is introduced to a decision-making system was analyzed theoretically and empirically. Previous work dealt with causality and responsibility in human-automation systems without considering the effects of how the automation's role is presented to users. An existing analytical model for predicting the human contribution to outcomes was adapted to accommodate the context of automation. An aided signal detection experiment with 400 participants was conducted to assess the correspondence of observed behavior to model predictions. The context in which the automation's role is presented affected users' tendency to follow its advice. When automation made decisions, and users only supervised it, they tended to contribute less to the outcome than in systems where the automation had an advisory capacity. The adapted theoretical model for human contribution was generally aligned with participants' behavior. The specific way automation is integrated into a system affects its use and the perceptions of user involvement, possibly altering overall system performance. The research can help design systems with automation-assisted decision-making and provide information on regulatory requirements and operational processes for such systems.

Small unmanned helicopter system identification based on the weighted least square method and improved grey wolf optimisation algorithm

Abstract Aiming to address the issue of low accuracy in model predictions obtained from fitting frequency domain response curves for small unmanned helicopters during the process of modeling their flight dynamics, this study proposes a system identification algorithm based on the combination of weighted least squares and improved grey wolf optimisation algorithm. The algorithm utilises the weighted least squares method to obtain the initial model structure, optimises the initial model parameters using the improved grey wolf optimisation algorithm, and enhances the local search and global optimisation ability of the grey wolf optimisation algorithm by introducing an improved grey wolf subgrouping rule, nonlinear convergence factor and dynamic cooperative rule. Ultimately, this approach establishes a dynamic model for small, unmanned helicopters. The identified model is validated using flight test data, with findings demonstrating that this method achieves higher accuracy in model identification and better fits to frequency domain response curves, thus providing a more accurate reflection of the flight dynamics of small unmanned helicopters.

Race to the Bottom: Competition and Quality in Science

Abstract This paper investigates how competition to publish first and thereby establish priority impacts the quality of scientific research. We begin by developing a model where scientists decide whether and how long to work on a given project. When deciding how long they should let their projects mature, scientists trade off the marginal benefit of higher quality research against the marginal risk of being preempted. Projects with the highest scientific potential are the most competitive because they induce the most entry. Therefore, the model predicts these projects are also the most rushed and lowest quality. We test the predictions of this model in the field of structural biology using data from the Protein Data Bank (PDB), a repository for structures of large macromolecules. An important feature of the PDB is that it assigns objective measures of scientific quality to each structure. As suggested by the model, we find that structures with higher ex-ante potential generate more competition, are completed faster, and are lower quality. Consistent with the model, and with a causal interpretation of our empirical results, these relationships are mitigated when we focus on structures deposited by scientists who — by nature of their employment position — are less focused on publication and priority. We estimate that the costs associated with improving these low-quality structures are between 1.5 and 8.8 billion dollars since the PDB’s founding in 1971.

A numerical model for duricrust formation by water table fluctuations

Abstract. Duricrusts are hard mineral layers forming in climatically contrasted environments. They form in tropical to arid environments and can be currently observed all around the world in areas such as Europe, Africa, South America, India, and Australia. In most cases, they cap hills and appear to protect softer layers beneath. Two main hypotheses have been proposed for the formation of duricrusts; i.e. the hydrological or transported model, where the enrichment in the hardening element (iron for ferricretes, silica for silcretes, or calcium carbonates for calcretes) is the product of leaching and precipitation through fluctuations in the water table during contrasted seasonal cycles, and the laterization or in situ model, where the formation of duricrusts is the final compacting stage of laterization. In this article, we present the first numerical geomorphological model for the formation of duricrusts based on the hydrological hypothesis. The model is an extension to an existing regolith formation model, where the position of the water table is used to predict the formation of a hardened layer at a rate set by a characteristic timescale, τ, and over a depth set by the range of fluctuations in the water table, λ. Hardening causes a decrease in surface erodibility, which we introduce in the model as a dimensionless factor, κ, that multiplies the surface transport coefficient of the model. Using the model, we show under which circumstances duricrusts form by introducing two dimensionless numbers that combine the model parameters (λ and τ), as well as parameters representing external forcing like precipitation rate and uplift rate. We demonstrate that when using model parameter values obtained by independent constraints from field observations, hydrology, and geochronology, the model predictions reproduce the observed conditions for duricrust formation. We also show that a strong feedback exists due to duricrust formation on the shape of the regolith and the position of the water table. Finally, we demonstrate that although duricrusts protect elements of the landscape, their efficiency in doing so is significantly lower than their inherent strength.

A model for effective conductivity of polymer nanocomposites containing MXene nanosheets

AbstractThis paper introduces a groundbreaking model to evaluate the conductivity of nanocomposites comprising MXene nanosheets. The model simulates the effective conductivity considering MXene dimensions, MXene volume fraction, interphase thickness, percolation threshold, contact distance, and tunneling resistance. The model's predictions align well with empirical conductivity results obtained various laboratory samples. The scrutiny of elements impacting effective conductivity is affirmed, given the assumption of contact resistance and the operation of the MXene/interphase network. Slender MXene nanosheets and expansive contacts lead to an elevated level of effective conductivity. Moreover, the effective conductivity shows a direct correlation with the MXene loading, while a higher percolation onset produces a poorer conductivity. Based on the model's outputs, an insulative nanocomposite is identified via the thinnest interphase ( &lt; 1 nm), the thickest MXene (t &gt; 4 nm), the smallest MXene volume fraction ( &lt; 0.01), and the lowest percentage of networked nanosheets ( &lt; 0.05). Contrariwise, the most remarkable conductivity as 25.6 S/m is attained by the thinnest MXene nanosheets (t = 1 nm). In addition, the narrowest tunnels (tunneling distance of 1 nm) yield the uppermost effective conductivity of 6.2 S/m in the system.Highlights This study proposes a model for conductivity of polymer MXene nanocomposites. MXene size, interphase depth, contact distance, and tunneling resistance are considered. The predictions agree with the experimental conductivity data of several samples. A higher conductivity is obtained by the bigger contact area and thicker interphase. The narrowest tunnels (1 nm) produce the uppermost effective conductivity of 6.2 S/m.

The stability of anisotropic compact stars influenced by dark matter under teleparallel gravity: an extended gravitational deformation approach

In our investigation, we pioneer the development of geometrically deformed strange stars within the framework of teleparallel gravity theory through gravitational decoupling via the complete geometric deformation (CGD) technique. The significant finding is the precise solution for deformed strange star (SS) models achieved through the vanishing complexity factor scenario. Further, we introduce the concept of space-time deformation caused by dark matter (DM) content in DM haloes, leading to perturbations in the metric potentials gtt and grr components. Mathematically, this DM-induced deformation is achieved through the CGD method, where the decoupling parameter α governs the extent of DM influence. To validate our findings, we compare our model predictions with observational constraints, including GW190814 (with a mass range of 2.5-2.67M⊙) and neutron stars (NSTRs) such as EXO 1785-248 [mass=1.3-0.2+0.2M⊙], 4U 1608-52 [mass=1.74-0.14+0.14M⊙], and PSR J0952-0607 [mass=2.35-0.17+0.17M⊙]. Our investigation delves into the stability of the model by considering causality conditions, Herrera’s cracking method, the adiabatic index, and the Harrison–Zeldovich–Novikov criterion. We demonstrate that the developed model mimics a wide range of recently observed pulsars. To emphasize its compatibility, we highlight the predicted mass and radius in tabular form by varying both the parameters α and ζ1. Notably, our findings are consistent with the observation of gravitational waves from the first binary merger event. Furthermore, we compare our results with those obtained for a slow-rotating configuration. In addition to this, we discuss the moment of inertia using the Bejger–Haensel approach in this formulation.

A Comparative Study of Fine-Tuning Deep Learning Models for Leaf Disease Identification and Classification

Innovative agricultural solutions are needed to detect and classify leaf diseases early across crop species and environments. This study compares deep learning approaches, focusing on Convolutional Neural Networks (CNN) and Vision Transformers (VTs), to identify leaf diseases early and accurately for scalable crop management and productivity. Optimizing CNNs, Explainable Transfer Learning (ETPLDNet) using ResNet50 architecture, and LEViT leaf disease diagnosis are compared. The CNN model, optimized with dynamic hyperparameters, achieved an impressive 99.58% accuracy for leaf disease classification, demonstrating its effectiveness in feature extraction and classification precision. On the other hand, the VT-based LEViT model, which leverages self-attention mechanisms and Explainable AI (XAI), achieved 95.22% accuracy but offers enhanced interpretability and generalization capabilities due to its transformer-based architecture. This distinction illustrates that while CNNs excel in accuracy, VTs provide a more transparent decision-making process and better handle the complex variances in plant leaf diseases, making them ideal for precision agriculture. The combined use of CNNs and VTs showcases the strengths of each model, with CNN focusing on high classification precision and VTs offering improved interpretability and adaptability for various leaf disease conditions. The use of XAI enables the models to highlight important areas in plant leaf images that influence the model's decisions, offering a transparent and interpretable decision-making process that allows researchers and farmers to understand why a particular diagnosis or classification was made. This ability to visualize and explain the reasoning behind the model predictions is crucial to increasing trust in AI-driven solutions in agriculture. By combining the high precision of CNN and the interpretability of VT with XAI, this study offers a robust approach to improving crop disease management and precision agriculture.

Online Payment Fraud Detection Model Using Machine Learning

The dependability and safety of UPI-based financial transactions are vital for maintaining public confidence, facilitating seamless digital payments, and ensuring economic stability. This paper introduces a software model aimed at identifying and predicting fraudulent behavior in UPI transactions through the use of advanced technologies, which encompass machine learning algorithms, real-time monitoring, and behavioral analysis. The Use of the software is to reduce financial risks by implementing a fraud assessment system that relies on historical transaction data and current information. By utilizing machine learning algorithms along with rule-based methods, the system is able to recognize abnormal transaction patterns, evaluate user behavior, and identify irregularities. The Prediction of model find the risk of transactions based on factors such as the amount, frequency, location, and behavioral patterns of users to highlight potentially suspicious activities. The suggested system functions by implementing four main phases: user engagement, analysis of historical data, continuous monitoring of real-time data, and fraud prediction analysis. This methodology produces comprehensive reports on the risks associated with transactions and offers proactive measures to identify and thwart fraudulent actions within UPI transactions Key Words UPI Transactions, Fraud Detection, Machine Learning, Real-Time Monitoring, Behavioral Analysis, Risk Assessment,

Enhancing Prediction of Electron Affinity and Ionization Energy in Liquid Organic Electrolytes for Lithium-Ion Batteries Using Machine Learning

The electron affinity (EA) and ionization energy (IE) of battery electrolyte molecules are pivotal in determining the electrochemical properties of electrode materials, significantly impacting the performance and safety of batteries. In this study, we employed a suite of machine learning algorithms, including linear regression, gradient boosting, CatBoost, XGBoost, and random forest regression, to predict these properties. Our analysis leveraged a comprehensive dataset of 24432 electrolyte molecules extracted from the Joint Center for Energy Storage Research (JCESR) Molecules database, and used RDKit and Mordred tools to generate a series of molecular descriptors. Through meticulous feature engineering and selection using the maximum relevance minimum redundancy (mRMR) algorithm, we identified the most influential molecular features for predicting EA and IE. The CatBoost model emerged as the most accurate, outperforming other models with its ability to handle complex nonlinear relationships and provide robust predictions with lower errors. The model's predictions were further validated through 10-fold cross-validation, demonstrating its generalization capability and resistance to overfitting. The CatBoost model achieved a root mean squared error (RMSE) of 0.239 eV for IE and 0.427 eV for EA, with R-squared (R2) values of 0.944 and 0.879, respectively. SHapley Additive exPlanations (SHAP) value analysis elucidated the contribution of each feature to the model's predictions, highlighting the molecular characteristics that significantly influence EA and IE. Our findings offer valuable insights for the design of novel electrolyte materials with tailored electronic properties, advancing the development of high-performance battery technologies.

Deep learning assisted prediction of osteogenic capability of orthopedic implant surfaces based on early cell morphology.

The surface modification of titanium (Ti) and its alloys is crucial for improving their osteogenic capability, as their bio-inert nature limits effective osseointegration despite their prevalent use in orthopedic implants. However, these modification methods produce varied surface properties, making it challenging to standardize criteria for assessing the osteogenic capacity of implant surfaces. Additionally, traditional evaluation experiments are time-consuming and inefficient. To overcome these limitations, this study introduced a high-throughput, efficient screening method for assessing the osteogenic capability of implant surfaces based on early cell morphology and deep learning. The Orthopedic Implants-Osteogenic Differentiation Network (OIODNet) was developed using early cell morphology images and corresponding alkaline phosphatase (ALP) activity values from cells cultured on Ti and its alloy surfaces, achieving performance metrics exceeding 0.98 across all six evaluation parameters. Validation through metal-polyphenol network (MPN) coatings and cell experiments demonstrated a strong correlation between OIODNet's predictions and actual ALP activity outcomes, confirming its accuracy in predicting osteogenic potential based on early cell morphology. The Osteogenic Predictor application offers an intuitive tool for predicting the osteogenic capacity of implant surfaces. Overall, this research highlights the potential to accelerate progress at the intersection of artificial intelligence and biomaterials, paving the way for more efficient screening of osteogenic capabilities in orthopedic implants. STATEMENT OF SIGNIFICANCE: By leveraging deep learning, this study introduces the Orthopedic Implants-Osteogenic Differentiation Network (OIODNet), which utilizes early cell morphology data and alkaline phosphatase (ALP) activity values to provide a high-throughput, accurate method for predicting osteogenic capability. With performance metrics exceeding 0.98, OIODNet's accuracy was further validated through experiments involving metal-polyphenol network (MPN) coatings, showing a strong correlation between the model's predictions and experimental outcomes. This research offers a powerful tool for more efficient screening of implant surfaces, marking a transformative step in the integration of artificial intelligence and biomaterials, while opening new avenues for advancing orthopedic implant technologies.

Overall effects of topsoil elements on cancer mortality in rural Greece: A modeling approach.

This ecological study examines cancer mortality rates in 61 rural Greek municipalities, covering in total 7,305,554 person-years from 2000 to 2015, based on the Hellenic Statistical Authority data. Topsoil concentrations of Mn, Ni, Pb, Be, As and Cd in Greek grazing land samples were obtained from the GEMAS (Geochemical Mapping of Agricultural and Grazing land Soil) project. Municipalities of rural regions with population of up to 20,000 people were selected as the study area and were divided into four quartiles, according to their age-specific cancer mortality rates, to identify the most divergent areas of low/high mortality. Fusion of demographic data with topsoil measurements aimed to reveal their potential association. Towards a univariate analysis, comparing each element individually across the low/high mortality regions revealed significant overlaps in distributions of concentrations for all elements. Machine learning multi-element models were built to predict whether an area would have a low or high cancer mortality rate, based simultaneously and solely on the concentrations of the six elements of the region, for each group. The out-of-sample median accuracy of the model was very high (83.3%) for the 0-29 age group, high (66.7%) for the 30-39, 40-49 and 50-59 age groups, while the accuracy decreased significantly above age of 60. The interaction of Mn, As, Ni, and partially Be, contributes most to the model's prediction for the 0-29 age group, As becomes the main contributor in the 30-39 age group, while all six elements contribute almost equally for the 40-49 group, and the interaction of Ni, Mn and Be becomes important in the 50-59 age group. The findings of this study strongly suggest that topsoil elements may jointly influence the frequency and geographical distribution of cancer.

Smartphone pupillometry with machine learning differentiates ischemic from hemorrhagic stroke: A pilot study.

Similarities between acute ischemic and hemorrhagic stroke make diagnosis and triage challenging. We studied a smartphone-based quantitative pupillometer for differentiation of acute ischemic and hemorrhagic stroke. Stroke patients were recruited prior to surgical or interventional treatment. Smartphone pupillometry was used to quantify components of the pupillary light reflex (PLR). A synthetic minority oversampling technique (SMOTE) was applied to correct sample size imbalance. Four binary classification model types were trained using all possible combinations of the PLR components with 10-fold cross validation stratified by cohort. Models were evaluated for accuracy, sensitivity, specificity, area under the curve (AUC), and F1 score. The three best-performing models were selected based on AUC. Shapley additive explanation plots were produced to explain PLR parameter impacts on model predictions. Eleven subjects with intraparenchymal hemorrhage and 22 subjects with acute ischemic stroke were enrolled. One way ANOVA demonstrated significant differences between healthy control data, AIS, and IPH in five out of seven PLR parameters. After SMOTE, each class had n=22 PLR recordings for model training. The best-performing model was random forest using a combination of latency, mean and maximum constriction velocity, and mean dilation velocity to discriminate between stroke types with 91.5% (95% confidence interval: 84.1-98.9) accuracy, 90% (82.9-97.1) sensitivity, 93.3% (83-100) specificity, 0.917 (0.847-0.987) AUC, and 90.7% (84.1-97.3) F1 score. Smartphone-based quantitative pupillometry could be useful in differentiating between acute ischemic and hemorrhagic stroke.

Robust EKF based on the framework of dynamic data reconciliation for state estimation of chemical processes with gaussian/non-gaussian measurement noise

State estimation plays a critical role in modern industry. The extended Kalman filter (EKF) is effective for nonlinear chemical processes with Gaussian noise, but it struggles when gross errors are present. This paper addresses this limitation by reformulating the EKF within the dynamic data reconciliation (DDR) framework, resulting in a robust DDR-based EKF tailored for state estimation in chemical processes with non-Gaussian measurement noise. The combination of random and gross errors is modeled using a contaminated Gaussian distribution. Model predictions are incorporated as prior knowledge, and a fixed-point iterative strategy is employed to update the posterior probability. Additionally, a first-order linearization technique is applied for convergence analysis. The robustness and effectiveness of the DDR-based EKF are demonstrated through both a classic mathematical example and a styrene polymerization reaction. Simulation results show that the DDR-based EKF effectively mitigates gross errors, achieving reliable state estimation.

The sedimentation probability model based on the particle size distribution and cake morphology in the cross-flow membrane process

Evolution of non-uniform cake structure in the cross-flow membrane process was investigated in this study. Considering the real cake surface morphology and particle size distribution, the sedimentation probability of foulant (γ) was calculated, then a new flux prediction model based on the mass conservation on the cake surface was proposed. Results showed that the predictions of the new model had excellent agreements with experimental data (R2 > 0.97), wherein cake surface heights obeyed normal distribution and protrusion heights approximately obeyed exponential distribution. Meanwhile, γ decreased with cross-flow velocity from 4.59 % (0.24 m/s) to 2.53 % (0.48 m/s), whereas γ increased with trans-membrane pressure from 2.80 % (0.1 MPa) to 6.44 % (0.2 MPa). In addition, a skin layer which was approximately about 20 % of total cake thickness but possessed 70 % of the total resistance, was also observed in the cross-flow mode. These results provided a brand-new perspective (tuning cake morphology or structure) on the essence of membrane fouling alleviation.

Establishment of strength evolution model and life prediction for diatomite-modified concrete under freeze-thaw conditions

Establishment of strength evolution model and life prediction for diatomite-modified concrete under freeze-thaw conditions