Advanced Machine Learning Methods Research Articles

Advanced tree-based machine learning methods for predicting the seismic response of regular and irregular RC frames

In this study, predicting the maximum drift ratio (MDR) for two types of structures: 8-story regular (R8) and vertically irregular (IR8) reinforced concrete (RC) frames are focused. To accomplish this, advanced tree-based machine learning (ML) methods, namely Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Stochastic Gradient Boosting (SGB) are utilized. 2300 ground motion (GM) records, includes twenty-one input parameters and their nonlinear time history analyses results are considered as a dataset. The performance of the RF, XGBoost, and SGB models was examined considering the impact of the data pre-processing issue to test whether the use of data pre-processing approaches in model development could improve the prediction performance of the advanced ML models. Various data pre-processing techniques, including outlier detection using boxplot, missing data imputation using RF, data transformation using log transformation, and feature selection using Hill Climbing are employed as part of ML pipeline. The models' performance is evaluated using six different well-established statistical evaluation metrics. Subsequently, the results are systematically compared to identify the most effective prediction model. Our findings highlight the substantial impact of data pre-processing on the performance of ML models when predicting seismic responses, specifically the MDR of RC frames. Notably, RF consistently outperformed other ML models for both R8 and IR8 structures in the case of the simple ML pipeline strategy and data pre-processing approach. The RF model achieved R2 scores of R2 = 0.87 and R2 = 0.91 on the test data before implementing data pre-processing for R8 and IR8, respectively. After data pre-processing, R2 scores showed an improvement of 12% and 6.22% in the cases of R8 and IR8. In contrast, the performance scores of SGB consistently fell behind those of RF and XGBoost models.

Enhancing Bioactive Compound Classification through the Synergy of Fourier-Transform Infrared Spectroscopy and Advanced Machine Learning Methods

Bacterial infections and resistance to antibiotic drugs represent the highest challenges to public health. The search for new and promising compounds with anti-bacterial activity is a very urgent matter. To promote the development of platforms enabling the discovery of compounds with anti-bacterial activity, Fourier-Transform Mid-Infrared (FT-MIR) spectroscopy coupled with machine learning algorithms was used to predict the impact of compounds extracted from Cynara cardunculus against Escherichia coli. According to the plant tissues (seeds, dry and fresh leaves, and flowers) and the solvents used (ethanol, methanol, acetone, ethyl acetate, and water), compounds with different compositions concerning the phenol content and antioxidant and antimicrobial activities were obtained. A principal component analysis of the spectra allowed us to discriminate compounds that inhibited E. coli growth according to the conventional assay. The supervised classification models enabled the prediction of the compounds’ impact on E. coli growth, showing the following values for accuracy: 94% for partial least squares-discriminant analysis; 89% for support vector machine; 72% for k-nearest neighbors; and 100% for a backpropagation network. According to the results, the integration of FT-MIR spectroscopy with machine learning presents a high potential to promote the discovery of new compounds with antibacterial activity, thereby streamlining the drug exploratory process.

Predicting Indian GDP with Machine Learning: A Comparison of Regression Models

In recent years, the application of machine learning algorithms in economic forecasting has gained significant traction. This study focuses on predicting the Gross Domestic Product (GDP) of India by harnessing advanced machine learning techniques. We assembled a comprehensive dataset incorporating time series analysis and inflation rates from diverse sources. Our investigation involved a comparative analysis employing both linear and polynomial regression methods to discern the most accurate predictive model. Our findings underscore the superiority of the polynomial regression model, which excelled in capturing non-linear relationships between independent variables and GDP. Specifically, the polynomial regression model achieved an impressive prediction accuracy rate of surpassing the accuracy achieved by the linear regression model. This study underscores the critical role of advanced machine learning algorithms in economic forecasting. By emphasizing the significance of high-quality datasets and the application of techniques like polynomial regression, we illustrate how these factors can substantially enhance the precision of economic forecasts. These insights hold substantial implications for policy makers and businesses alike, aiding informed decision-making and effective economic policy development. This study serves as a valuable reference for future research, highlighting the importance of advanced machine learning methods and robust data sources in economic forecasting.

Pre-hospital glycemia as a biomarker for in-hospital all-cause mortality in diabetic patients - a pilot study

BackgroundType 2 Diabetes Mellitus (T2DM) presents a significant healthcare challenge, with considerable economic ramifications. While blood glucose management and long-term metabolic target setting for home care and outpatient treatment follow established procedures, the approach for short-term targets during hospitalization varies due to a lack of clinical consensus. Our study aims to elucidate the impact of pre-hospitalization and intra-hospitalization glycemic indexes on in-hospital survival rates in individuals with T2DM, addressing this notable gap in the current literature.MethodsIn this pilot study involving 120 hospitalized diabetic patients, we used advanced machine learning and classical statistical methods to identify variables for predicting hospitalization outcomes. We first developed a 30-day mortality risk classifier leveraging AdaBoost-FAS, a state-of-the-art ensemble machine learning method for tabular data. We then analyzed the feature relevance to identify the key predictive variables among the glycemic and routine clinical variables the model bases its predictions on. Next, we conducted detailed statistical analyses to shed light on the relationship between such variables and mortality risk. Finally, based on such analyses, we introduced a novel index, the ratio of intra-hospital glycemic variability to pre-hospitalization glycemic mean, to better characterize and stratify the diabetic population.ResultsOur findings underscore the importance of personalized approaches to glycemic management during hospitalization. The introduced index, alongside advanced predictive modeling, provides valuable insights for optimizing patient care. In particular, together with in-hospital glycemic variability, it is able to discriminate between patients with higher and lower mortality rates, highlighting the importance of tightly controlling not only pre-hospital but also in-hospital glycemic levels.ConclusionsDespite the pilot nature and modest sample size, this study marks the beginning of exploration into personalized glycemic control for hospitalized patients with T2DM. Pre-hospital blood glucose levels and related variables derived from it can serve as biomarkers for all-cause mortality during hospitalization.

Performance Scrutiny of Price Prediction on Blockchain Technology Using Machine Learning

Objective: This study, aiming to assess the accurateness of machine learning algorithms for the prediction of cryptocurrency prices, using blockchain technology, was conducted to attest to its effectiveness. The main goal is to evaluate the precision, solidity, and prospects of such models under variable and fragmented trading premises. Methodology: Dataset containing historical crypto prices, transaction volumes, emotions of investors, etc. is formed from different and various sources. Different machine learning approaches including linear regression, support vector machine, random forest, and deep learning networks will be used to construct the prediction model. The split of the dataset into training and testing subsets is performed to better evaluate and estimate the models' performance. Feature engineering techniques and parameter tuning are included in the process of making models more precise. Models are the ones which are the trained using historical data and then later tested out on the unseen data for the purpose of accurate results and generalization power. Results and Discussion: From the experiment it is evident that a shift of different machine learning algorithms exhibits differing outcomes. While certain models can yield admirable results in forecasting crypto prices in shorter terms, however, there are some models that are not capable of capturing the complicated impurities. Factors like data quality, function selection and model complexity are at the core of predictive power. The learning networks demonstrated the capability to reproduce nonlinearity and temporal relatedness in cryptocurrency value data. Conclusion: Cryptocurrency forecasting through blockchain technology involves the use of machine learning techniques which has a lot of potential. The success of these models also depends on the quality of the data, feature refinement and model choice. Research progress should focus on refining currently existing models, considering new data inputs, and introduction of advanced machine learning methods which should be able to increase forecast accuracy in rapidly changing cryptocurrency markets.

Automated Structure Discovery for Scanning Tunneling Microscopy.

Scanning tunneling microscopy (STM) with a functionalized tip apex reveals the geometric and electronic structures of a sample within the same experiment. However, the complex nature of the signal makes images difficult to interpret and has so far limited most research to planar samples with a known chemical composition. Here, we present automated structure discovery for STM (ASD-STM), a machine learning tool for predicting the atomic structure directly from an STM image, by building upon successful methods for structure discovery in noncontact atomic force microscopy (nc-AFM). We apply the method on various organic molecules and achieve good accuracy on structure predictions and chemical identification on a qualitative level while highlighting future development requirements for ASD-STM. This method is directly applicable to experimental STM images of organic molecules, making structure discovery available for a wider scanning probe microscopy audience outside of nc-AFM. This work also allows more advanced machine learning methods to be developed for STM structure discovery.

Energy-saving effect assessment of various factors in container plant factories: A data-driven random forest approach

Plant factory is one of the controlled environment agriculture forms with huge potential to alleviate food crisis, but the high energy cost restricts its widespread adoption. Numerous researches have explored various factors for energy-saving in plant factories in their settings, but there is a lack of analysis of the importance of these factors in energy saving. In this work, the energy-saving effect assessment of various factors in the container plant factory is investigated. Four cities (Harbin, Taiyuan, Shanghai, and Guangzhou), three plant densities (cultivation area: floor area=100 %, 150 %, and 200 %), and two temperature/humidity setpoints (20/22 ℃, 60/70 %, and 16/22 ℃, 50/95 %) are selected as operating conditions to cover different weather conditions and plant heat loads. The energy-saving effect of each factor is calculated using a random forest algorithm based on large amounts of energy simulation data. We identify that envelope overall heat transfer coefficient (U), air conditioner coefficient of performance (COP), and light efficacy (Efficacy) are three factors that have the largest impact on energy-saving in plant factories, in which light efficacy is the most important factor. Simultaneous optimization of these three factors could possibly reduce electricity consumption by ∼50 % compared to the base case. Finally, employing weight-light intensity correlation, the minimum specific energy consumption is approximately 4.76 kWh/kg lettuce fresh weight. This study utilizes advanced machine learning methods to sort out important factors and shows that significant energy reduction may be achieved by optimizing dominant factors, which gives a general guidance for future designers to build energy-efficient plant factories.

Extremely Low Frequency (ELF) Electromagnetic Signals as a Possible Precursory Warning of Incoming Seismic Activity

We analyzed a large number (77) of low-to-medium-magnitude earthquakes (M3.5–M6.5) that occurred within a period of three years (2020–2022) in the Southern half of Greece in relation to the ELF activity in that region and time period. In most cases, characteristic ELF signals appear up to 20 days before the earthquakes. This observation may add an important new element to the Lithospheric–Atmospheric–Ionospheric scenario, thus contributing to a better prediction of incoming earthquakes. We discuss the role of ELF observations in reliable seismic forecasting. We conclude that the magnitude of an earthquake larger than M4.0 and the distance of the epicenter shorter than 300 km from the recording site is needed for typical pre-seismic signals to be observed. Finally, we remark that a reliable prediction of earthquakes could result from an integrated project of multi-instrumental observations, where all the known variety of precursors would be included, and the whole data set would be analyzed by advanced machine learning methods.

Utilizing Machine Learning to Forecast Smoking Behavior

Abstract: The examination of smoking habits, extensively studied over time, has posed challenges in accurately predicting and thoroughly analyzing its determinants. Previous research efforts struggled to precisely predict smoking behavior due to the presence of continuous target variables, which hindered the application of crucial feature selection techniques like mutual information. This study aims to tackle these hurdles through an innovative approach that integrates data preprocessing, feature engineering, and advanced machine learning methods. To address the issue of continuous target variables, our methodology involves categorizing smoking behavior into discrete groups, enabling the utilization of feature selection techniques such as mutual information scores. Logistic regression, Gaussian Naive Bayes, and Random Forest Classifier models are utilized to achieve highly accurate predictions of smoking behavior. The Select KBest method is employed to evaluate the importance of features based on mutual information scores. The investigation delves into various health markers, including BMI, haemoglobin levels, and cholesterol, offering comprehensive insights into their influence on smoking habits. Furthermore, Principal Component Analysis (PCA) is implemented to reduce dimensionality while preserving essential information from the dataset. Through this novel approach and a steadfast commitment to ethical data collection practices, our objective is to advance the comprehension of smoking behavior, surmounting past challenges, and providing valuable insights for public health initiatives and smoking cessation endeavors. The paper assesses outcomes using specified algorithms and parameters, presenting a comparative analysis to enhance the clarity and reliability of our findings.

Model-agnostic explanations for survival prediction models.

Advanced machine learning methods capable of capturing complex and nonlinear relationships can be used in biomedical research to accurately predict time-to-event outcomes. However, these methods have been criticized as "black boxes" that are not interpretable and thus are difficult to trust in making important clinical decisions. Explainable machine learning proposes the use of model-agnostic explainers that can be applied to predictions from any complex model. These explainers describe how a patient's characteristics are contributing to their prediction, and thus provide insight into how the model is arriving at that prediction. The specific application of these explainers to survival prediction models can be used to obtain explanations for (i) survival predictions at particular follow-up times, and (ii) a patient's overall predicted survival curve. Here, we present a model-agnostic approach for obtaining these explanations from any survival prediction model. We extend the local interpretable model-agnostic explainer framework for classification outcomes to survival prediction models. Using simulated data, we assess the performance of the proposed approaches under various settings. We illustrate application of the new methodology using prostate cancer data.

Abstract 6217: Differential genome-wide gene regulatory networks for lung cancer

Abstract Background: Lung cancer remains the leading cause of cancer death worldwide due to the fact that most cases are diagnosed at distant stages, and treatment is less effective. During the past two decades, much effort has been devoted to understanding tumor biology and developing targeted therapeutics, which have significantly improved patient survival. With the increasingly available multi-omics data from healthy controls and lung cancer patients, we constructed gene regulatory networks for the two main subtypes of non-small-cell lung cancer (NSCLC), the most common type of lung cancer. In addition, differential networks were constructed by integrative analysis of data from lung cancer tissue and tissues from healthy lungs. Methods: Transcriptomics and genomic data were obtained from the patients with lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) tissue available in The Cancer Genome Atlas (TCGA) project, and both types of omics data of normal lung tissue were downloaded from The Genotype-Tissue Expression (GTEx) consortium. The networks for each subtype and the differential networks were constructed by utilizing our newly developed machine learning tools, two-stage penalized least square (2SPLS) approach, and analysis of variance of directed networks (NetANOVA), respectively. Results: For the top three identified subnetworks of LUAD, we identified some well-known genes associated with lung cancer, such as EGFR, IRF1, KRAS, ERBB2, PIK3CA, MYC, at bootstrap frequency >90%. Based on the top three subnetworks with a bootstrap frequency of 100%, results from the Ingenuity Pathway Analysis (IPA) indicated several significant pathways, primary immunodeficiency signaling and communication between innate and adaptive immune cells. GO enrichment analysis showed statistically significant biological processes, including immune response, extracellular matrix organization, and significant KEGG pathways such as Hematopoietic cell lineage. The differential networks between LUAD, LUSC, and healthy lungs showed the same regulations across all three groups, such as the regulation between EEF2 and RACK1, the perturbations between LUAD and healthy (two-way regulations between HIGD2A and ARL10), and LUSC and healthy (GRB14 regulates EIF3E) respectively. In addition, we identified the differences between LUAD and LUSC, such as bi-directional regulations between CLTB and ARL10. Conclusion: Although all groups share most regulations, we identified differential gene regulations between healthy lung tissue and lung cancer tissue and between LUAD and LUSC of NSCLC. Using multi-omics data generated from lung tissues, coupled with advanced machine learning methods, the causal relationships between genome-wide genes from the analysis will help us understand the molecular mechanism of lung cancer and facilitate the development of personalized treatment strategies. Citation Format: Min Zhang, Zhongli Jiang, Dabao Zhang. Differential genome-wide gene regulatory networks for lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6217.

Rent Price Prediction with Advanced Machine Learning Methods: A Comparison of California and Texas

The forecast of rent prices in dynamic housing markets is of fundamental importance to renters, landlords, investors, and politicians alike. Machine learning models offer flexibility, excel at modeling complex relationships, and provide outstanding forecast precision. This study compares advanced machine learning models, extreme gradient boosting regressor (XGBoost), light gradient boosting machine (LightGBM), random forest, ridge regression, and two ensemble approaches, to predict California and Texas rent prices. The two ensemble approaches include a hybrid regression of averaging base models and a 2-level stacked generalization model. The results revealed that a stacked generalization ensemble with base models random forest, XGBoost, LightGBM, and meta-model ridge regression achieved the best performance for the California dataset by generating the lowest MSE and highest R2 of 46116.3 and 0.8858, respectively. In contrast, random forest outperformed both ensemble models with the lowest MSE and MAPE of 18401.93 and 9.7003%, respectively, and the highest R2 of 0.7992. These methodologies can assess future rental property worth, serve as indicators for market dynamics, and aid in establishing real estate policies, thereby providing practical guidance to individuals, businesses, and policymakers.

Machine learning study using 2020 SDHS data to determine poverty determinants in Somalia

Extensive research has been conducted on poverty in developing countries using conventional regression analysis, which has limited prediction capability. This study aims to address this gap by applying advanced machine learning (ML) methods to predict poverty in Somalia. Utilizing data from the first-ever 2020 Somalia Demographic and Health Survey (SDHS), a cross-sectional study design is considered. ML methods, including random forest (RF), decision tree (DT), support vector machine (SVM), and logistic regression, are tested and applied using R software version 4.1.2, while conventional methods are analyzed using STATA version 17. Evaluation metrics, such as confusion matrix, accuracy, precision, sensitivity, specificity, recall, F1 score, and area under the receiver operating characteristic (AUROC), are employed to assess the performance of predictive models. The prevalence of poverty in Somalia is notable, with approximately seven out of ten Somalis living in poverty, making it one of the highest rates in the region. Among nomadic pastoralists, agro-pastoralists, and internally displaced persons (IDPs), the poverty average stands at 69%, while urban areas have a lower poverty rate of 60%. The accuracy of prediction ranged between 67.21% and 98.36% for the advanced ML methods, with the RF model demonstrating the best performance. The results reveal geographical region, household size, respondent age group, husband employment status, age of household head, and place of residence as the top six predictors of poverty in Somalia. The findings highlight the potential of ML methods to predict poverty and uncover hidden information that traditional statistical methods cannot detect, with the RF model identified as the best classifier for predicting poverty in Somalia.

Adaptive solar power generation forecasting using enhanced neural network with weather modulation

Solar power generation forecasting plays a vital role in optimizing grid management and stability, particularly in renewable energy-integrated power systems. This research paper presents a comprehensive study on solar power generation forecasting, evaluating traditional and advanced machine learning methods, including ARIMA, Exponential Smoothing, Support Vector Regression, Random Forest, Gradient Boosting, and Physics-based Models. Moreover, we propose an innovative Enhanced Artificial Neural Network (ANN) model, which incorporates Weather Modulation and Leveraging Prior Forecasts to enhance prediction accuracy. The proposed model is evaluated using real-world solar power generation data, and the results demonstrate its superior performance compared to traditional methods and other machine learning approaches. The Enhanced ANN model achieves an impressive Root Mean Square Error (RMSE) of 0.116 and a Mean Absolute Percentage Error (MAPE) of 36.26% . The integration of Weather Modulation allows the model to adapt to changing weather conditions, ensuring reliable forecasts even during adverse scenarios. Leveraging Prior Forecasts enables the model to capture short-term trends, reducing forecasting errors arising from abrupt weather changes. The proposed Enhanced ANN model showcases its potential as a promising tool for precise and reliable solar power generation forecasting, contributing to the efficient integration of solar energy into the power grid and advancing sustainable energy practices.

Predictive models for flexible pavement fatigue cracking based on machine learning

Pavement performance prediction is crucial for ensuring the longevity and safety of road networks. In our extensive study, we employ a diverse array of techniques to enhance fatigue performance models in flexible pavements. The methodology begins with Random Forest feature selection, identifying the top 15 critical variables that significantly impact pavement performance. These variables form the basis for subsequent model development. Our investigation into model performance indicates the superiority of advanced machine learning methods such as Regression Trees (RT), Gaussian Process Regression (GPR), Support Vector Machines (SVM), Ensemble Trees (ET), and Artificial Neural Networks (ANN) over traditional linear regression methods. This consistent outperformance underscores their potential to reshape forecasting accuracy. Through extensive model optimization, we reveal robust performance across both complete and selected feature sets, emphasizing the importance of meticulous feature selection in enhancing forecast accuracy. The accuracy of our best optimized machine learning model is highlighted by its Performance Measurement metrics: RMSE of 22.416, MSE of 502.46, R-squared of 0.80848, and MAE of 8.9958. Additionally, comparative analysis with previous empirical models demonstrates that our best optimized machine learning model outperforms existing empirical models. This work underscores the significance of feature curation in pavement performance prediction, highlighting the potential of sophisticated modeling methodologies. Embracing cutting-edge technologies facilitates data-driven decisions, ultimately contributing to the development of more robust road networks, enhancing safety, and prolonging lifespan.

Predicting Personal Exposure to PM2.5 Using Different Determinants and Machine Learning Algorithms in Two Megacities, China

The primary aim of this study is to explore the utility of machine learning algorithms for predicting personal PM2.5 exposures of elderly participants and to evaluate the effect of individual variables on model performance. Personal PM2.5 was measured on five consecutive days across seasons in 66 retired adults in Beijing (BJ) and Nanjing (NJ), China. The potential predictors were extracted from routine monitoring data (ambient PM2.5 concentrations and meteorological factors), basic questionnaires (personal and household characteristics), and time-activity diary (TAD). Prediction models were developed based on either traditional multiple linear regression (MLR) or five advanced machine learning methods. Our results revealed that personal PM2.5 exposures were well predicted by both MLR and machine learning models with predictors extracted from routine monitoring data, which was indicated by the high nested cross-validation (CV) R2 ranging from 0.76 to 0.88. The addition of predictors from either the questionnaire or TAD did not improve predictive accuracy for all algorithms. The ambient PM2.5 concentrations were the most important predictor. Overall, the random forest, support vector machine, and extreme gradient boosting algorithms outperformed the reference MLR method. Compared with the traditional MLR approach, the CV R2 of the RF model increased up to 7% (from 0.82±0.13 to 0.88±0.10), while the RMSE reduced up to 18% (from 19.8±5.4 to 16.3±4.5) in BJ.

Streamlining multi-hole probe calibration using artificial neural networks

Accurately measuring the three-dimensional flow field characteristics in complex flow fields, particularly in turbomachines, is of utmost importance and is commonly achieved through experimental methods. Multi-hole pressure probes are a proven measuring technique to provide precise readings of flow characteristics in both two-dimensional (2D) and three-dimensional (3D) flow fields. However, manufacturing imperfections have an impact on their measuring accuracy so each probe has a slightly different measuring behavior. In order to take these specific characteristics into account a thorough calibration process prior to utilization has to be applied. This calibration process involves positioning the probe in various flow conditions in a wind tunnel and collecting all relevant data. A model is then established to correlate the measured pressure distribution across the probe holes with known flow conditions. Unfortunately, this calibration process is time-consuming and requires costly equipment.To address these challenges, this study introduces a novel calibration technique that leverages advanced machine learning methods, specifically artificial neural networks (ANNs), to significantly streamline the process. ANNs are renowned for their ability to model diverse systems and adapt and generalize well, making them a promising solution for simplifying the calibration of multi-hole probes. In this research, an ANN calibration model was developed specifically for a 5-hole probe, and its configuration was optimized through an automated hyperparameter optimization process. The performance of the ANN model was then compared to that of a conventional polynomial model. After confirming its accuracy, the ANN model was successfully re-adapted and applied to new probes, requiring minimal calibration data collection. The results highlight the potential of ANNs in probe calibration, as they enable a simplified process that demands fewer calibration data points.

Analyzing chloride diffusion for durability predictions of concrete using contemporary machine learning strategies

Analyzing the migration of chloride ions in concrete can enhance its durability and decrease the probability of corrosion. Furthermore, machine learning (ML) offers a promising, cost-effective, and less intricate alternative to traditional experimental methods for accurately forecasting cementitious composites' chloride migration coefficient (Dnssm). This study utilized advanced ML methods like support vector machine (SVM), artificial neural network (ANN), and k-nearest neighbor (KNN) to forecast the Dnssm of concrete accurately. A comprehensive database was established considering significant input features that can impact Dnssm. The models undergo verification with unseen data, utilizing various statistical performance measures. The verification findings validate that all three models (SVM, KNN, ANN) accurately predicted the Dnssm with high precision. However, the SVM model outperformed with a higher coefficient of determination (R2) score of 0.90, whereas ANN and KNN yielded R2 scores of 0.88 and 0.83, respectively. The model has the capacity to substitute laborious, resource-intensive and time-consuming lab testing. Moreover, the RReliefF analysis indicated that water/binder and cement type are the prominent input factors affecting the target variable Dnssm.

SARS-CoV-2 Infection Impairs Oculomotor Functions: A Longitudinal Eye-tracking Study.

Although Severe Acute Respiratory Syndrome Coronavirus 2 infection (SARS-CoV-2) is primarily recognized as a respiratory disease, mounting evidence suggests that it may lead to neurological and cognitive impairments. The current study used three eye-tracking tasks (free-viewing, fixation, and smooth pursuit) to assess the oculomotor functions of mild infected cases over six months with symptomatic SARS-CoV-2 infected volunteers. Fifty symptomatic SARS-CoV-2 infected, and 24 self-reported healthy controls completed the eye-tracking tasks in an initial assessment. Then, 45, and 40 symptomatic SARS-CoV-2 infected completed the tasks at 2- and 6-months post-infection, respectively. In the initial assessment, symptomatic SARS-CoV-2 infected exhibited impairments in diverse eye movement metrics. Over the six months following infection, the infected reported overall improvement in health condition, except for self-perceived mental health. The eye movement patterns in the free-viewing task shifted toward a more focal processing mode and there was no significant improvement in fixation stability among the infected. A linear discriminant analysis shows that eye movement metrics could differentiate the infected from healthy controls with an accuracy of approximately 62%, even 6 months post-infection. These findings suggest that symptomatic SARSCoV- 2 infection may result in persistent impairments in oculomotor functions, and the employment of eye-tracking technology can offer valuable insights into both the immediate and long-term effects of SARS-CoV-2 infections. Future studies should employ a more balanced research design and leverage advanced machine-learning methods to comprehensively investigate the impact of SARSCoV- 2 infection on oculomotor functions.

From EU-SoilHydroGrids to HU-SoilHydroGrids: A leap forward in soil hydraulic mapping

Spatially explicit, quantitative information on soil hydraulic properties is required in various modelling schemes. At European scale, EU-SoilHydroGrids proved its applicability in a number of studies, in ecological predictions, geological and hydrological hazard assessment, agri-environmental models, among others. Inspired by its continental antecedent, an analogous, but larger scale, national, 3D soil hydraulic database was elaborated for the territory of Hungary (HU-SoilHydroGrids) supported by various improvements (i–iv) in the computation process. Pedotransfer functions (PTFs) were built in the form of i) advanced machine learning methods and ensemble models, and trained on the ii) national soil hydrophysical dataset. The set of predictors used in PTFs was supplemented by iii) additional environmental auxiliary variables. Spatial layers of the soil hydraulic parameters were generated using iv) 100 m resolution information on primary soil properties, namely DOSoReMI.hu. HU-SoilHydroGrids provides information on the most frequently required soil hydraulic properties (water content at saturation, field capacity and wilting point, saturated hydraulic conductivity and van Genuchten parameters for the description of the moisture retention curve) with national coverage at 100 m spatial resolution down to 2 m depth for six GSM standard depth layers. The HU-SoilHydroGrids has significantly lower squared error in the case of describing the moisture retention curve and hydraulic conductivity than the EU-SoilHydroGrids. The derived 3D soil hydraulic database (ver1.0) is presently available in National Laboratory for Water Science and Water Safety for project partners in order to test its functional performance in describing hydrological and ecological processes.