Robust Fitting Research Articles

Recognition of Conus species using a combined approach of supervised learning and deep learning-based feature extraction.

Cone snails are venomous marine gastropods comprising more than 950 species widely distributed across different habitats. Their conical shells are remarkably similar to those of other invertebrates in terms of color, pattern, and size. For these reasons, assigning taxonomic signatures to cone snail shells is a challenging task. In this report, we propose an ensemble learning strategy based on the combination of Random Forest (RF) and XGBoost (XGB) methods. We used 47,600 cone shell images of uniform size (224 x 224 pixels), which were split into an 80:20 train-test ratio. Prior to performing subsequent operations, these images were subjected to pre-processing and transformation. After applying a deep learning approach (Visual Geometry Group with a 16-layer deep model architecture) for feature extraction, model specificity was further assessed by including multiple related and unrelated seashell images. Both classifiers demonstrated comparable recognition ability on random test samples. The evaluation results suggested that RF outperformed XGB due to its high accuracy in recognizing Conus species, with an average precision of 95.78%. The area under the receiver operating characteristic curve was 0.99, indicating the model's optimal performance. The learning and validation curves also demonstrated a robust fit, with the training score reaching 1 and the validation score gradually increasing to 95 as more data was provided. These values indicate a well-trained model that generalizes effectively to validation data without significant overfitting. The gradual improvement in the validation score curve is crucial for ensuring model reliability and minimizing the risk of overfitting. Our findings revealed an interactive visualization. The performance of our proposed model suggests its potential for use with datasets of other mollusks, and optimal results may be achieved for their categorization and taxonomical characterization.

Using Differential Scanning Calorimetry to Measure the Energetic Properties of Residual Sludge and Catalysts from the Textile, Tannery, and Galvanic Industries

This research delved into the energetic properties of catalysts synthesized from residual sludge from the textile, galvanic, and tannery industries. The experimental process consisted of an initial heat treatment to activate their catalytic properties and a thermal analysis employing differential scanning calorimetry (DSC). This technique permitted the investigation of the materials’ thermal behavior as a function of temperature, ranging from 142 to 550 °C, effectively controlling the heating rates and pressure conditions. The data gathered were the input for constructing specific heat models through polynomial regression employing the least squares method. These models were subsequently used to estimate variations in the enthalpy and entropy for both the sludge and catalysts through integration. Third-degree polynomials primarily characterized the specific heat models that accurately represented the samples’ thermal behavior, considering variations in their physicochemical properties that influenced it. The catalysts derived from residual sludge from the textile industry exhibited the models with the most robust statistical fit. Concurrently, the catalysts from the galvanic industry displayed noteworthy similarities with the bibliographic data across various temperature points. The mathematical models determined the specific heat (Cp) as a function of temperature, which, in turn, was used to estimate the enthalpy and entropy variations in the sludge and catalysts under study. The highest enthalpy value corresponded to the sludge and catalyst obtained from the tannery industry, with a Cp of 5.60 J/g-K at 603 K and 2.45 J/g-K at 445.6 K. Finally, the third-degree polynomials showed the best mathematical models since (1) they considered the variations in the physicochemical properties that intervened in the behavior of Cp as a function of temperature; (2) they presented a better statistical fit; and (3) they showed consistency with the existing information in the literature for the textile industry and the galvanic industries.

Effect of technology multiplier: A framework for analysis of innovation perspectives on production segment allocation

In the realm of production systems, determining the optimal segment allocation remains a central concern. While several existing models address this issue, a significant gap remains as many overlook the critical role of innovation and lack a holistic perspective. This paper presents a model that emphasizes innovation capabilities and introduces the concept of a “Technology Multiplier” underscoring the compounding influence of technology and innovation on production segment allocation decisions. Within this work, we focus on preliminary studies to establish the “Technology Multiplier” concept employing an Analytical Hierarchy Process (AHP) with sensitivity analysis. The validity of our approach is demonstrated through four case studies from three industries, illustrating the relevance of our elaborated metrics for the concept of “Technology Multipliers”. In particular, a leading automotive company uses our findings to reach a more appropriate strategic decision aligned with innovation and production growth, compared to its previous decisions. These results not only demonstrate a robust fit with our proposed metrics but also indicate that our framework lays the foundation for further research on the “Technology Multiplier”, enriching the decision-making process for production segment allocation.

Analysing the impact of age, wealth, mother's education, area of residence, and gender on child weight: a regression approach

Abstract The purpose of this article is to analyse the impact of age, wealth, mother's education, area of residence, and gender all together on child weight using multiple linear regression model. Previous studies have examined the effects of these factors on child weight singly or in pairs but not all together. The various factors affecting child weight that contribute to the broader issues of malnutrition and obesity need to be examined. This study addresses this issue with the aim of identifying the most significant predictors of healthy child weight. The data were collected from a secondary source for a sample size of 760 children aged 0-59 months. The data were analysed using multiple linear regression with SPSS software. The correlation coefficient (0.861) between the observed and predicted values of the dependent variable (child weight) suggests a strong positive correlation. 74.2 per cent of the variability in child weight is explained by the predictors in this model, indicating a good fit. The model also explains 74per cent of the variance, suggesting a robust fit. Age, gender, wealth, and mother's education are significant predictors of child weight, whilst area of residence is not statistically significant, though location-specific strategies may still be beneficial for child health outcomes. Based on the findings of this study, public health programs should prioritise age-appropriate nutrition, gender-sensitive interventions, and enhanced maternal education to improve child weight outcomes.

A framework for automated supraglacial lake detection and depth retrieval in ICESat-2 photon data across the Greenland and Antarctic ice sheets

Abstract. Water depths of supraglacial lakes on the ice sheets are difficult to monitor continuously due the lakes' ephemeral nature and inaccessible locations. Supraglacial lakes have been linked to ice shelf collapse in Antarctica and accelerated flow of grounded ice in Greenland. However, the impact of supraglacial lakes on ice dynamics has not been quantified accurately enough to predict their contribution to future mass loss and sea level rise. This is largely because ice-sheet-wide assessments of meltwater volumes rely on models that are poorly constrained due to a lack of accurate depth measurements. Various recent case studies have demonstrated that accurate supraglacial lake depths can be obtained from NASA's Ice, Cloud and land Elevation Satellite (ICESat-2) ATL03 photon-level data product. ATL03 comprises hundreds of terabytes of unstructured point cloud data, which has made it challenging to use this bathymetric capability at scale. Here, we present two new algorithms – Flat Lake and Underlying Ice Detection (FLUID) and Surface Removal and Robust Fit (SuRRF) – which together provide a fully automated and scalable method for lake detection and along-track depth determination from ATL03 data and establish a framework for its large-scale implementation using distributed high-throughput computing. We report FLUID–SuRRF algorithm performance over two regions known to have significant surface melt – central West Greenland and the Amery Ice Shelf catchment in East Antarctica – during two melt seasons. FLUID–SuRRF reveals a total of 1249 ICESat-2 lake segments up to 25 m deep, with more water during higher-melt years. In the absence of ground-truth data, manual annotation of test data suggests that our method reliably detects melt lakes along ICESat-2's ground tracks whenever the lake bed is visible or partially visible and estimates water depths with a mean absolute error <0.27 m. These results imply that our proposed framework has the potential to generate a comprehensive data product of accurate meltwater depths across both ice sheets.

Redefining the impact of professional development in education with ProDES (Professional Development Evaluation Scale)

This study introduces the Professional Development Evaluation Scale (ProDES), a tool that has been developed to evaluate the impact of professional development as it relates to participants' Learning and Use of New Knowledge and Skills, Organisational Support, Student Learning Outcomes, and reactions. Grounded in Guskey’s (2000) framework for evaluating Professional Development, ProDES was developed with data from five study groups in Turkey and underwent refinement across four factors. Exploratory and Confirmatory Factor Analyses confirmed the scale's structure, accounting for 62.72% of the total variance, with robust fit indices. Within this, ProDES demonstrated high internal consistency and test-retest reliability, with significant correlations validating its effectiveness. The scale's high internal consistency and test-retest reliability ensure that it can be used to make evidence-informed decisions that can foster more effective and supportive professional development activities. As a result, by identifying which professional development initiatives lead to improvements, those associated with professional development can use resources more efficiently, leading to enhanced school and system-wide improvements. Moreover, the use of ProDES can also help schools and education systems track progress over time, making ProDES an invaluable tool for continuous improvement and strategic planning.

Robust Estimation for Threshold Autoregressive Moving-Average Models

Threshold autoregressive moving-average (TARMA) models extend the popular TAR model and are among the few parametric time series specifications to include a moving average in a nonlinear setting. The state dependent reactions to shocks is particularly appealing in Economics and Finance. However, no theory is currently available when the data present heavy tails or anomalous observations. Here we provide the first theoretical framework for robust M-estimation for TARMA models and study its practical relevance. Under mild conditions, we show that the robust estimator for the threshold parameter is super-consistent, while the estimators for autoregressive and moving-average parameters are strongly consistent and asymptotically normal. The Monte Carlo study shows that the M-estimator is superior, in terms of both bias and variance, to the least squares estimator, which can be heavily affected by outliers. The findings suggest that robust M-estimation should be generally preferred to the least squares method. We apply our methodology to a set of commodity price time series; the robust TARMA fit presents smaller standard errors and superior forecasting accuracy. The results support the hypothesis of a two-regime non-linearity characterized by slow expansions and fast contractions.

Analytical model for reactive contaminant back-diffusion from low-permeability aquitards with internal source zones

Back diffusion of pollutants from low-permeability aquitards to adjacent high-permeability aquifers is a key issue that hinders site remediation. Previous back diffusion studies have typically assumed that a non-reactive contaminant source zone is located only in the aquifer, with scant attention paid to situations where the source zone resides in aquitards with degradation capacity. This study develops a novel back-diffusion analytical model for rapidly predicting reactive contaminant release from aquitards with internal source zones. The model incorporates the dynamic dissolution of internal sources in a multilayer aquitard with contaminant diffusion, adsorption, and degradation. General solutions were obtained utilizing the Laplace transform, Fourier finite cosine transform, variable substitution, and global matrix methods, validated against experimental data and numerical outcomes. Results reveal that previous back-diffusion models tend to underestimate the spatial extent and depth of aquitard contamination as well as aquifer plume tailing risk when the contaminant source is located within the aquitards. Based on numerous simulations and robust data fitting, the predictive functions for the back-diffusion plume tailing time regarding geometry and release parameters of internal sources were proposed, highly facilitating the risk evaluation of low-permeability sites. Furthermore, accelerating degradation reactions and expanding the degradation zone could efficiently mitigate contaminant back-diffusion from aquitards, implying that active bio-barriers are promising for remediating low-permeability soils.

Predicting corrosion current density in magnesium alloy battery anodes: machine learning approach using rapid miner

ABSTRACT Magnesium (Mg) alloys are increasingly recognised for their potential as anode materials in batteries due to their high specific capacity and cost-effectiveness. However, their susceptibility to corrosion poses a significant challenge to practical applications. This study explores the effectiveness of various machine learning models – local polynomial regression, Decision Trees, Random Forest, Gradient Boost, and Extreme Gradient Boost – in predicting the corrosion behaviour of Mg alloy battery anodes in a 3.5 wt.% NaCl electrolyte. Experimental data on the corrosion current density of Mg alloys are collected and analysed using the RapidMiner tool for model training and evaluation. The findings highlight that the Extreme Gradient Boost (XGBoost) model excels among other machine learning techniques in accurately forecasting the corrosion rates of Mg alloys. XGBoost achieves an exceptional R2 value of 0.9931 in Tafel plots, indicating a robust fit to the data as evidenced by scatter plots. Furthermore, the XGBoost model demonstrates strong positive monotonic and ordinal relationships between actual and predicted corrosion current densities, as indicated by a Spearman correlation coefficient of 1 and a Kendall tau coefficient of 0.99. These outcomes underscore XGBoost's potential as a powerful tool for enhancing the understanding and management of corrosion in Mg alloy battery applications.

Outlier detection in cardiac diffusion tensor imaging: Shot rejection or robust fitting?

Cardiac diffusion tensor imaging (cDTI) is highly prone to image corruption, yet robust-fitting methods are rarely used. Single voxel outlier detection (SVOD) can overlook corruptions that are visually obvious, perhaps causing reluctance to replace whole-image shot-rejection (SR) despite its own deficiencies. SVOD’s deficiencies may be relatively unimportant: corrupted signals that are not statistical outliers may not be detrimental. Multiple voxel outlier detection (MVOD), using a local myocardial neighbourhood, may overcome the shared deficiencies of SR and SVOD for cDTI while keeping the benefits of both. Here, robust fitting methods using M-estimators are derived for both non-linear least squares and weighted least squares fitting, and outlier detection is applied using (i) SVOD; and (ii) SVOD and MVOD. These methods, along with non-robust fitting with/without SR, are applied to cDTI datasets from healthy volunteers and hypertrophic cardiomyopathy patients. Robust fitting methods produce larger group differences with more statistical significance for MD, FA, and E2A, versus non-robust methods, with MVOD giving the largest group differences for MD and FA. Visual analysis demonstrates the superiority of robust-fitting methods over SR, especially when it is difficult to partition the images into good and bad sets. Synthetic experiments confirm that MVOD gives lower root-mean-square-error than SVOD.

TREND OF CASSAVA YIELD/PRODCUTION IN IN KOGI EAST SENATORIAL DISTRICT, KOGI STATE, NIGERIA

This study examined the trends in cassava yield and production in the Kogi East Senatorial District of Kogi State, Nigeria, focusing on the perceptions of local farmers regarding yield changes over the past five years and the factors influencing these trends. A total of 403 respondents were surveyed, revealing that 71.3% of the farmers perceived an increase in cassava yields, while 28.7% reported stable yields. The primary factors contributing to the perceived increase included the use of improved cassava varieties (58.1%) and climate variability (41.9%). Furthermore, the statistical analysis indicated a stationary R-squared value of 0.792, suggesting a robust model fit for predicting future production trends. Forecasts indicated a steady increase in cassava production per hectare from 2024 to 2033, with projected yields rising from 17.33 tons per hectare in 2024 to 18.54 tons per hectare by 2033. Despite these positive trends, the study identified significant challenges posed by climate variability, with 76.7% of respondents recognizing a high impact on production. The findings underscore the urgent need for targeted interventions, including the promotion of climate-resilient cassava varieties, improved access to quality planting materials, and the implementation of sustainable agricultural practices. Overall, this study provides critical insights into the cassava production landscape in Kogi East and highlights the importance of collaborative efforts among stakeholders to enhance productivity and ensure food security in the region.

Design, synthesis, and evaluation of a pyrazole-based corrosion inhibitor: a computational and experimental study

By employing a synergistic blend of experimental and theoretical methodologies, we investigated the corrosion inhibition efficacy of a synthesized pyrazole derivative (BM-01) in a solution of hydrochloric acid (1 M). We utilized molecular dynamics (MD) simulations, scanning electron microscopy (SEM), density functional theory (DFT), complexation, plus electrochemical impedance spectroscopy (EIS). We conducted weight loss (WL) measurements from 298 to 328 K. Inhibition efficacy reached a maximum at a BM-01 concentration of 10−3 M, achieving 90.0% (EIS), 90.40% (WL), and 90.38% (potentiodynamic polarization (PDP)). SEM unveiled the shielding of the carbon-steel surface from acid-induced damage by BM-01. The Langmuir adsorption isotherm exhibited a robust fit with a low sum of squares, standard deviation, and a high correlation coefficient. PDP findings indicated that BM-01 acted as a mixed-type inhibitor, predominantly favoring the cathodic process, suggesting potential corrosion-mitigation properties. Theoretical analyses involving DFT, MD simulations, and radial distribution function were conducted to postulate a mechanism and identify an inhibitory layer. Theoretical outcomes aligned closely with experimental data, thereby reinforcing the validity of our findings.

Role of AI-Driven Personalized E-Learning In Enhancing Economic Competitiveness: A Comparative Analysis of Developed And Developing Countries

This study explores the influence of AI-driven personalized e-learning on the Global Competitiveness Index (GCI) across six countries, comprising three developed (USA, Germany, Japan) and three developing (Pakistan, India and Nigeria) nations, over a span of 17 years (2007-2023). Utilizing a least squares regression model, the analysis incorporates key variables including the Artificial Intelligence Global Index (AIGI), Human Development Index (HDI), and R&D expenditure as control variables to assess their impact on GCI. The model demonstrates a robust fit, with an R2 value of 96.79% and an adjusted R2 value of 96.50%, indicating that AI-enhanced e-learning is a significant predictor as independent variables of economic competitiveness GCI. The findings underscore the significant role of investment in AI and e-learning technologies, particularly in enhancing economic competitiveness, as evidenced by strong positive correlations between GCI and AIGI, HDI, and R&D expenditure. The results are consistent with IBM's research, which highlights the importance of these variables in driving national competitiveness in the global market. The study concludes that AI-driven personalized e-learning is a critical lever for improving GCI, particularly for developing countries aiming to close the competitiveness gap with developed nations. The implications of this research suggest that targeted investments in AI and e-learning can substantially contribute to sustainable economic growth and international competitiveness. This study offers valuable insights for policymakers and educational institutions aiming to enhance their nation's competitiveness through strategic investments in AI-driven educational technologies.

On sea ice emission modeling for MOSAiC's L-band radiometric measurements

Abstract The retrieval of sea ice thickness using L-band passive remote sensing requires robust models for emission from sea ice. In this work, measurements obtained from surface-based radiometers during the MOSAiC expedition are assessed with the Burke, Wilheit and SMRT radiative transfer models. These models encompass distinct methodologies: radiative transfer with/without wave coherence effects, and with/without scattering. Before running these emission models, the sea ice growth is simulated using the Cumulative Freezing Degree Days (CFDD) model to further compute the evolution of the ice structure during each period. Ice coring profiles done near the instruments are used to obtain the initial state of the computation, along with Digital Thermistor Chain (DTC) data to derive the sea ice temperature during the analyzed periods. The results suggest that the coherent approach used in the Wilheit model results in a better agreement with the horizontal polarization of the in situ measured brightness temperature. The Burke and SMRT incoherent models offer a more robust fit for the vertical component. These models are almost equivalent since the scattering considered in SMRT can be safely neglected at this low frequency, but the Burke model misses an important contribution from the snow layer above sea ice. The results also suggest that a more realistic permittivity falls between the spheres and random needles formulations, with potential for refinement, particularly for L-band applications, through future field measurements.

Mathematical modeling of ions adsorption from water/wastewater sources via porous materials: A machine learning-based approach

This paper developed the predictive modeling of substance concentration (C) utilizing the input parameters x and y, for analysis of adsorption process. Employing three distinct machine learning models—Multilayer Perceptron (MLP), polynomial regression (PR), and Support Vector Machine (SVM)—the study investigates the efficacy of models in capturing the relationships between the inputs and output. The models are trained from data obtained from mass transfer calculations for removal of solute from solution via porous adsorbent. Furthermore, the hyper-parameters for each model are optimized through the utilization of the Political Optimizer (PO). The Multilayer Perceptron model emerges as a standout performer, showcasing an exceptional R-squared score of 0.9981, indicative of a robust fit to the data. Complemented by impressively low MAE and MSE values (7.94043E-01 and 2.0420E+00, respectively), the MLP model attests to its ability to provide accurate predictions and discern underlying patterns within the dataset. The polynomial regression model, while slightly trailing behind the MLP in terms of R-squared score (0.95929), revealed commendable predictive performance. Support Vector Machine also proves to be a formidable contender, boasting a robust R-squared score of 0.96055.

Translation and validation of the abbreviated Prefrontal Symptoms Inventory (PSI-20): A tool for assessing prefrontal symptoms in English-speaking populations.

This study introduces the translation and validation of the Prefrontal Symptoms Inventory (PSI) into English, aiming to provide an ecologically valid tool for assessing prefrontal symptoms in English-speaking populations in the United States. The prefrontal cortex (PFC) plays a crucial role in executive functions and other higher-order cognitive processes, with dysfunctions in this area associated with various cognitive, emotional and behavioural changes. Despite the existence of established tools like the Dysexecutive Questionnaire (DEX), the PSI addresses limitations found in the literature, presenting a novel ecologically valid tool for assessing prefrontal symptoms. The current study, involving 226 English-speaking participants, lays a foundational step for validating the PSI for use in a new population. Semi-confirmatory factorial analysis revealed a unidimensional structure, mirroring the Spanish version with robust fit indicators. Additionally, in assessing convergent validity, the abbreviated version (PSI-20) exhibited high correlations with DEX scores and moderate correlations with Psychological Stress Scale and General Health Questionnaire-12 scores. These findings align with previous reports, supporting the PSI-20's measurement of similar constructs related to prefrontal cortex activity and mental health components. The results of this study overall highlight the PSI's potential contribution to advancing prefrontal symptom evaluation in clinical and non-clinical settings.

Robust, fast and accurate mapping of diffusional mean kurtosis

Diffusional kurtosis imaging (DKI) is a methodology for measuring the extent of non-Gaussian diffusion in biological tissue, which has shown great promise in clinical diagnosis, treatment planning, and monitoring of many neurological diseases and disorders. However, robust, fast, and accurate estimation of kurtosis from clinically feasible data acquisitions remains a challenge. In this study, we first outline a new accurate approach of estimating mean kurtosis via the sub-diffusion mathematical framework. Crucially, this extension of the conventional DKI overcomes the limitation on the maximum b-value of the latter. Kurtosis and diffusivity can now be simply computed as functions of the sub-diffusion model parameters. Second, we propose a new fast and robust fitting procedure to estimate the sub-diffusion model parameters using two diffusion times without increasing acquisition time as for the conventional DKI. Third, our sub-diffusion-based kurtosis mapping method is evaluated using both simulations and the Connectome 1.0 human brain data. Exquisite tissue contrast is achieved even when the diffusion encoded data is collected in only minutes. In summary, our findings suggest robust, fast, and accurate estimation of mean kurtosis can be realised within a clinically feasible diffusion-weighted magnetic resonance imaging data acquisition time.

Robust, fast and accurate mapping of diffusional mean kurtosis.

Diffusional kurtosis imaging (DKI) is a methodology for measuring the extent of non-Gaussian diffusion in biological tissue, which has shown great promise in clinical diagnosis, treatment planning, and monitoring of many neurological diseases and disorders. However, robust, fast, and accurate estimation of kurtosis from clinically feasible data acquisitions remains a challenge. In this study, we first outline a new accurate approach of estimating mean kurtosis via the sub-diffusion mathematical framework. Crucially, this extension of the conventional DKI overcomes the limitation on the maximum b-value of the latter. Kurtosis and diffusivity can now be simply computed as functions of the sub-diffusion model parameters. Second, we propose a new fast and robust fitting procedure to estimate the sub-diffusion model parameters using two diffusion times without increasing acquisition time as for the conventional DKI. Third, our sub-diffusion-based kurtosis mapping method is evaluated using both simulations and the Connectome 1.0 human brain data. Exquisite tissue contrast is achieved even when the diffusion encoded data is collected in only minutes. In summary, our findings suggest robust, fast, and accurate estimation of mean kurtosis can be realised within a clinically feasible diffusion-weighted magnetic resonance imaging data acquisition time.

Study on Cycle Performance and Rate Performance of Lithium Cobalt Oxide

Lithium cobalt oxide (LiCoO2 ) cathode material, with its high energy density and operating voltage, is currently a mainstream material for lithium-ion battery cathodes.   However,  the  crystal  structure  collapse  and  lattice  oxygen  evolution under high-voltage conditions lead to rapid capacity decay, severely limiting its practical applications.   This paper  analyzes the main factors affecting the cycle performance and rate performance of lithium cobalt oxide, considering the physic- ochemical properties of the particles, including elemental content and particle size, and provides a mathematical model linking these properties to electrochemical per- formance.  The study offers insights for the practical production of high-voltage lithium cobalt oxide materials. At the beginning, we embarked on investigating the correlation between the physicochemical attributes (encompassing elemental composition and particle size) of lithium cobalt oxide and its cycle performance. An Ordinary Least Squares (OLS) linear regression model was formulated, yielding a robust fit with an R-Squared value of 0.94. This model was subsequently optimized through the application of an XGBoost algorithm, achieving an R-Squared value nearing unity, signifying a remarkable enhancement in model accuracy. Visual analysis of the results pinpointed the primary determinants of cycle performance, arranged in descending order of significance: 'Cycle Index,' Mg content, particle size distribution (D10), Zn, and Al. Then, our attention shifted to examining the link between the aforementioned physicochemical characteristics and the rate performance of lithium cobalt oxide. The Jarque-Bera and Shapiro-Wilk tests confirmed the normality of the data, fulfilling the prerequisites for hypothesis testing. An OLS linear regression model was developed, demonstrating a strong goodness-of-fit with an R-Squared value exceeding 0.8. This model was further honed with an XGBoost model, which achieved an R-Squared score approaching 1, indicating a substantial refinement in model precision. The visualization of model outcomes illuminated the key factors influencing rate performance, ranked in descending order of importance: particle size distribution (D50), Al, Mn, Zn, Mg, Ni, and Fe. Lastly, we devised an optimal strategy that encompasses the incorporation of strategic doping elements, the utilization of high-temperature sol-gel methods to bolster cycle performance, and coating modifications aimed at enhancing rate performance. This holistic approach fosters the structural stability of lithium cobalt oxide crystals, fortifying their high-potential cycle capabilities, and concurrently elevating their rate performance.

Mapping Multiphase Metals in Star-forming Galaxies: A Spatially Resolved UV+Optical Study of NGC 5253

We present a pioneering, spatially resolved, multiphase gas abundance study on the blue compact dwarf galaxy NGC 5253, targeting 10 star-forming (SF) clusters inside six far-UV (FUV) Hubble Space Telescope/Cosmic Origins Spectrograph (COS) pointings with cospatial optical Very Large Telescope/MUSE observations throughout the galaxy. The SF regions span a wide range of ages (1–15 Myr) and are distributed at different radii (50–230 pc). We performed a robust absorption-line profile fitting on the COS spectra, covering 1065–1430 Å in the FUV, allowing an accurate computation of neutral-gas abundances for 13 different ions sampling eight elements. These values were then compared with the ionized-gas abundances, measured using the direct method on MUSE integrated spectra inside analog COS apertures. Our multiphase, spatially resolved comparisons find abundances, which are lower in the neutral gas than the ionized gas by 0.22, 0.80, and 0.58 dex for log(O/H), log(N/H), and log(N/O), respectively. We modeled the chemical abundance distributions and evaluated correlations as a function of the radius and age. It was found that, while N, O, and N/O abundances decrease as a function of age in the ionized gas, they increase with age in the neutral gas. No strong correlations for N, O, or N/O were observed as a function of the radius. The N/O and N/H offsets between the phases were found to decrease with age, providing evidence that chemical enrichment happens differentially, first in the ionized-gas phase around 2–5 Myrs (due to N-rich Wolf–Rayet stars) and then mixing out into the cold neutral gas on longer timescales of 10–15 Myr.