Best-performing Model Research Articles

Attention-Driven Transfer Learning Model for Improved IoT Intrusion Detection

The proliferation of Internet of Things (IoT) devices has become inevitable in contemporary life, significantly affecting myriad applications. Nevertheless, the pervasive use of heterogeneous IoT gadgets introduces vulnerabilities to malicious cyber-attacks, resulting in data breaches that jeopardize the network’s integrity and resilience. This study proposes an Intrusion Detection System (IDS) for IoT environments that leverages Transfer Learning (TL) and the Convolutional Block Attention Module (CBAM). We extensively evaluate four prominent pre-trained models, each integrated with an independent CBAM at the uppermost layer. Our methodology is validated using the BoT-IoT dataset, which undergoes preprocessing to rectify the imbalanced data distribution, eliminate redundancy, and reduce dimensionality. Subsequently, the tabular dataset is transformed into RGB images to enhance the interpretation of complex patterns. Our evaluation results demonstrate that integrating TL models with the CBAM significantly improves classification accuracy and reduces false-positive rates. Additionally, to further enhance the system performance, we employ an Ensemble Learning (EL) technique to aggregate predictions from the two best-performing models. The final findings prove that our TL-CBAM-EL model achieves superior performance, attaining an accuracy of 99.93% as well as high recall, precision, and F1-score. Henceforth, the proposed IDS is a robust and efficient solution for securing IoT networks.

Comparative Analysis of Machine Learning Techniques for Water Consumption Prediction: A Case Study from Kocaeli Province.

This study presents a comparative analysis of various Machine Learning (ML) techniques for predicting water consumption using a comprehensive dataset from Kocaeli Province, Turkey. Accurate prediction of water consumption is crucial for effective water resource management and planning, especially considering the significant impact of the COVID-19 pandemic on water usage patterns. A total of four ML models, Artificial Neural Networks (ANN), Random Forest (RF), Support Vector Machines (SVM), and Gradient Boosting Machines (GBM), were evaluated. Additionally, optimization techniques such as Particle Swarm Optimization (PSO) and the Second-Order Optimization (SOO) Levenberg-Marquardt (LM) algorithm were employed to enhance the performance of the ML models. These models incorporate historical data from previous months to enhance model accuracy and generalizability, allowing for robust predictions that account for both short-term fluctuations and long-term trends. The performance of each model was assessed using cross-validation. The R2 and correlation values obtained in this study for the best-performing models are highlighted in the results section. For instance, the GBM model achieved an R2 value of 0.881, indicating a strong capability in capturing the underlying patterns in the data. This study is one of the first to conduct a comprehensive analysis of water consumption prediction using machine learning algorithms on a large-scale dataset of 5000 subscribers, including the unique conditions imposed by the COVID-19 pandemic. The results highlight the strengths and limitations of each technique, providing insights into their applicability for water consumption prediction. This study aims to enhance the understanding of ML applications in water management and offers practical recommendations for future research and implementation.

Prediction of aeration performance of different types of piano key weirs using different machine learning models

ABSTRACT Aeration is a cost-effective and efficient method for increasing the available oxygen or dissolved oxygen content in water bodies, which is crucial for the existence of aquatic life. However, conventional techniques for estimating aeration in different hydraulic structures are time-consuming and incorrect ways to approximate aeration. Therefore, new, computationally more efficient, and more accurate methods are required. In this article, three machine learning models are presented: (1) ELM (extreme learning machine) model, (2) online sequential extreme learning machine model, and (3) I-ELM (incremental extreme learning machine) model. These models assess the air conditioning capacity of the three variants of Piano Key Weirs (PKWs), denoted as A, B, and C, about Cd, Cs, and Cu, which are the three most important parameters for aeration efficiency at different temperatures. The model performance is evaluated and compared based on mean squared error, root-mean-square error, correlation coefficient, mean absolute error, and Nash–Sutcliffe efficiency. This research concludes that I-ELM is the best-performing model for complete available data that are time invariant.

Assessing plant pigmentation impacts: A novel approach integrating UAV and multispectral data to analyze atrazine metabolite effects from soil contamination

The objective of this study was to evaluate the levels of desethylatrazine (DEA), a hydrophilic metabolite of atrazine, and its impact on plant health. This was achieved by utilizing multispectral imagery captured by Unmanned Aerial Vehicles (UAVs) in combination with ground-measured data to assess photosynthetic pigment levels in Green Cos lettuce following atrazine application in agricultural soil. Strong correlations were found between DEA levels and chlorophyll a, chlorophyll b, and anthocyanin levels in lettuce (R² > 0.70), while the correlation with carotenoid levels was weaker (R² = 0.55). This disruption to the pigments could interfere with photosynthesis, potentially hindering the plant's growth and development, and ultimately leading to a reduction in yield. The Anthocyanin Reflectance Index (ARI) demonstrated a robust positive correlation with DEA, whereas the Normalized Difference Red Edge (NDRE), Leaf Chlorophyll Index (LCI), and Normalized Difference Vegetation Index (NDVI) displayed pronounced negative correlations. Incorporating ARI, LCI, and NDRE, with or without NDVI, provided the most accurate prediction of DEA levels, with an R² exceeding 0.96. NDRE emerged as the most efficient index for forecasting chlorophyll a and chlorophyll b levels. Modified Chlorophyll Absorption in Reflectance Index (MCARI) demonstrated the best fit for carotenoids, while ARI performed exceptionally well in describing actual measurements of anthocyanins (R² = 0.90). The best-performing VI models, developed from the selection of effective single variables, exhibited the best fit to actual pigment measurements (R² > 0.83). These findings underscore the role of UAV-derived multispectral imagery in assessing DEA levels and improving environmental monitoring, aiding in better planning for agriculture and environmental remediation to enhance ecosystem health and resilience.

Development and calibration of a mathematical model of HIV outcomes among Rwandan adults: informing equitable achievement of targets in Rwanda.

We developed and calibrated the Central Africa-International epidemiology Databases to Evaluate AIDS (CA-IeDEA) HIV policy model to inform equitable achievement of global goals, overall and across sub-populations, in Rwanda. We created a deterministic dynamic model to project adult HIV epidemic and care continuum outcomes, overall and for 25 subpopulations (age group, sex, HIV acquisition risk, urbanicity). Data came from the Rwanda cohort of CA-IeDEA, 2004-2020; Rwanda Demographic and Health Surveys, 2005, 2010, 2015; Rwanda Population-based HIV Impact Assessment, 2019; and the literature and reports. We calibrated the model to 47 targets by selecting the 50 best-fitting parameter sets among 20,000 simulations. Calibration targets reflected epidemic (HIV prevalence, incidence), global goals (percentage on antiretroviral therapy (ART) among diagnosed, percentage virally suppressed among on ART) and other (number on ART, percentage virally suppressed) indicators, overall and by sex. Best-fitting sets minimized the summed absolute value of the percentage deviation (AVPD) between model projections and calibration targets. Good model performance was mean AVPD ≤5% across the 50 best-fitting sets and/or projections within the target confidence intervals; acceptable was mean AVPD >5% and ≤15%. Across indicators, 1,841 of 2,350 (78.3%) model projections were a good or acceptable fit to calibration targets. For HIV epidemic indicators, 256 of 300 (85.3%) projections were a good fit to targets, with the model performing better for women (83.3% a good fit) than for men (71.7% a good fit). For global goals indicators, 96 of 100 (96.0%) projections were a good fit; model performance was similar for women and men. For other indicators, 653 of 950 (68.7%) projections were a good or acceptable fit. Fit was better for women than for men (percentage virally suppressed only) and when restricting targets for number on ART to 2013 and beyond. The CA-IeDEA HIV policy model fits historical data and can inform policy solutions for equitably achieving global goals to end the HIV epidemic in Rwanda. High-quality, unbiased population-based data, as well as novel approaches that account for calibration target quality, are critical to ongoing use of mathematical models for programmatic planning.

Exoplanet Detection Using Machine Learning : A Comparative Study Using Kepler Mission Data

The search for habitable planets outside our solar system has captivated scientists throughout the centuries. Discovery and characterization of exoplanets have been one of the most important endeavors of modern astronomy. With various space missions, we have significantly expanded our observational capacity, resulting in an abundance of information about the universe. The influx of more data necessitates the development of techniques that can aid astronomers in processing all the information more efficiently and in an automated manner. Machine learning in recent years has become an indispensable paradigm to automate complex tasks that are possible only by humans. This work explores the application of machine learning to detect exoplanets from NASA’s Kepler mission. Our dataset comprises Kepler Objects of Interest (KOIs), encompassing their characteristic features and confirmed exoplanet status. We experiment with multiple supervised classification techniques including classical, tree-based, and neural methods. The best-performing model Histogram Gradient Boosting achieves a strong performance of 94.6% precision and 94.1% recall on a held-out dataset demonstrating the strong potential of integrating machine learning techniques into astronomy, potentially leading to new insights into planetary systems outside the solar system.

The application of chemical similarity measures in an unconventional modeling framework c-RASAR along with dimensionality reduction techniques to a representative hepatotoxicity dataset

With the exponential progress in the field of cheminformatics, the conventional modeling approaches have so far been to employ supervised and unsupervised machine learning (ML) and deep learning models, utilizing the standard molecular descriptors, which represent the structural, physicochemical, and electronic properties of a particular compound. Deviating from the conventional approach, in this investigation, we have employed the classification Read-Across Structure-Activity Relationship (c-RASAR), which involves the amalgamation of the concepts of classification-based quantitative structure-activity relationship (QSAR) and Read-Across to incorporate Read-Across-derived similarity and error-based descriptors into a statistical and machine learning modeling framework. ML models developed from these RASAR descriptors use similarity-based information from the close source neighbors of a particular query compound. We have employed different classification modeling algorithms on the selected QSAR and RASAR descriptors to develop predictive models for efficient prediction of query compounds’ hepatotoxicity. The predictivity of each of these models was evaluated on a large number of test set compounds. The best-performing model was also used to screen a true external data set. The concepts of explainable AI (XAI) coupled with Read-Across were used to interpret the contributions of the RASAR descriptors in the best c-RASAR model and to explain the chemical diversity in the dataset. The application of various unsupervised dimensionality reduction techniques like t-SNE and UMAP and the supervised ARKA framework showed the usefulness of the RASAR descriptors over the selected QSAR descriptors in their ability to group similar compounds, enhancing the modelability of the dataset and efficiently identifying activity cliffs. Furthermore, the activity cliffs were also identified from Read-Across by observing the nature of compounds constituting the nearest neighbors for a particular query compound. On comparing our simple linear c-RASAR model with the previously reported models developed using the same dataset derived from the US FDA Orange Book (https://www.accessdata.fda.gov/scripts/cder/ob/index.cfm), it was observed that our model is simple, reproducible, transferable, and highly predictive. The performance of the LDA c-RASAR model on the true external set supersedes that of the previously reported work. Therefore, the present simple LDA c-RASAR model can efficiently be used to predict the hepatotoxicity of query chemicals.

Temporal trends in asteroid behaviour: a machine learning and N-body integration approach

ABSTRACT Asteroids pose significant threats to Earth, necessitating early detection for potential deflection. Leveraging machine learning (ML), we classify asteroids into near-Earth Asteroids (particularly Atens, Amors, Apollos, and Apoheles) and non-near-Earth asteroids, further categorizing them based on hazard potential. Training the seven models on a comprehensive data set of 4687 asteroids, we achieve high accuracy in prediction. The predictive capability of these models is critical for informed decision-making in planetary defense strategies. We apply different regularization techniques to prevent overfitting and validate the models using a large unseen data set. A rigorous long-term N-body integration spanning 1 Myr is executed utilizing the Mercury N-body integrator to illuminate the evolution of asteroid properties over extended temporal scales. Following this integration process, the best-performing ML model is employed to classify asteroids based on their orbital characteristics and hazardous status respectively. Our findings highlight the effectiveness of ML in asteroid classification and prediction, paving the way for large-scale applications. By dividing a 1 Myr integration into intervals, we uncover temporal trends in asteroid behaviour, revealing insights into hazard evolution and ejection patterns. Notably, initially, hazardous asteroids tend to transition to non-hazardous states over time, elucidating key dynamics in planetary defense. We illustrate these findings through plotted graphs, providing valuable insights into asteroid dynamics. These insights are instrumental in advancing our understanding of long-term asteroid behaviour, with significant implications for future research and planetary protection efforts.

Lung Cancer Staging Using Chest CT and FDG PET/CT Free-Text Reports: Comparison Among Three ChatGPT Large-Language Models and Six Human Readers of Varying Experience.

Background: Although radiology reports are commonly used for lung cancer staging, this task can be challenging given radiologists' variable reporting styles as well as reports' potentially ambiguous and/or incomplete staging-related information. Objective: To compare performance of ChatGPT large-language models (LLMs) and human readers of varying experience in lung cancer staging using chest CT and FDG PET/CT free-text reports. Methods: This retrospective study included 700 patients (mean age, 73.8±29.5 years; 509 male, 191 female) from four institutions in Korea who underwent chest CT or FDG PET/CT for non-small cell lung cancer initial staging from January, 2020 to December, 2023. Examinations' reports used a free-text format, written exclusively in English or in mixed English and Korean. Two thoracic radiologists in consensus determined the overall stage group (IA, IB, IIA, IIB, IIIA, IIIB, IIIC, IVA, IVB) for each report using the AJCC 8th-edition staging system, establishing the reference standard. Three ChatGPT models (GPT-4o, GPT-4, GPT-3.5) determined an overall stage group for each report using a script-based application programming interface, zero-shot learning, and prompt incorporating a staging system summary. Six human readers (two fellowship-trained radiologists with lesser experience than the radiologists who determined the reference standard, two fellows, two residents) also independently determined overall stage groups. GPT-4o's overall accuracy for determining the correct stage among the nine groups was compared with that of the other LLMs and human readers using McNemar tests. Results: GPT-4o had an overall staging accuracy of 74.1%, significantly better than the accuracy of GPT-4 (70.1%, p=.02), GPT-3.5 (57.4%, p<.001), and resident 2 (65.7%, p<.001); significantly worse than the accuracy of fellowship-trained radiologist 1 (82.3%, p<.001) and fellowship-trained radiologist 2 (85.4%, p<.001); and not significantly different from the accuracy of fellow 1 (77.7%, p=.09), fellow 2 (75.6%, p=.53), and resident 1 (72.3%, p=.42). Conclusions: The best-performing model, GPT-4o, showed no significant difference in staging accuracy versus fellows, but significantly worse performance versus fellowship-trained radiologists. The findings do not support use of LLMs for lung cancer staging in place of expert healthcare professionals. Clinical Impact: The findings indicate the importance of domain expertise for performing complex specialized tasks such as cancer staging.

Development of a Diagnostic Model for Pancreatic Ductal Adenocarcinoma Using Machine Learning and Blood-Based miRNAs

Introduction: Pancreatic ductal adenocarcinoma (PDAC) has the lowest survival rate among all major cancers due to a lack of symptoms in early stages, early detection tools, and optimal therapies for late-stage patients. Thus, effective and non-invasive diagnostic tests are greatly needed. Recently, circulating miRNAs have been reported to be altered in PDAC. They are promising biomarkers because of stability in the blood, ease of non-invasive detection, and convenient screening methods. This study aimed to use blood-based miRNA biomarkers and various analysis methods in the development of a machine-learning (ML) model for PDAC. Methods: Blood-based miRNAs associated with PDAC were collected from open sources. miRNA sequences, targeted genes, and involved pathways were used to construct a set of descriptors for an ML model. Results: Bioinformatics analysis revealed that most genes in pancreatic cancer and insulin signaling pathways were targeted by the PDAC-related miRNAs. The best-performing ML model with the Random Forest classifier was able to achieve an accuracy of 88.4%. Model evaluations of an independent PDAC-associated miRNAs test set had 100% accuracy while non-cancer miRNAs had 52.4% accuracy, indicating specificity to PDAC. Conclusions: Our results suggest an ML model developed using blood-based miRNA biomarkers’ target gene, pathway, and sequence features could be potentially implicated in PDAC diagnostics.

A machine learning approach towards assessing consistency and reproducibility: an application to graft survival across three kidney transplantation eras.

In South Africa, between 1966 and 2014, there were three kidney transplant eras defined by evolving access to certain immunosuppressive therapies defined as Pre-CYA (before availability of cyclosporine), CYA (when cyclosporine became available), and New-Gen (availability of tacrolimus and mycophenolic acid). As such, factors influencing kidney graft failure may vary across these eras. Therefore, evaluating the consistency and reproducibility of models developed to study these variations using machine learning (ML) algorithms could enhance our understanding of post-transplant graft survival dynamics across these three eras. This study explored the effectiveness of nine ML algorithms in predicting 10-year graft survival across the three eras. We developed and internally validated these algorithms using data spanning the specified eras. The predictive performance of these algorithms was assessed using the area under the curve (AUC) of the receiver operating characteristics curve (ROC), supported by other evaluation metrics. We employed local interpretable model-agnostic explanations to provide detailed interpretations of individual model predictions and used permutation importance to assess global feature importance across each era. Overall, the proportion of graft failure decreased from 41.5% in the Pre-CYA era to 15.1% in the New-Gen era. Our best-performing model across the three eras demonstrated high predictive accuracy. Notably, the ensemble models, particularly the Extra Trees model, emerged as standout performers, consistently achieving high AUC scores of 0.95, 0.95, and 0.97 across the eras. This indicates that the models achieved high consistency and reproducibility in predicting graft survival outcomes. Among the features evaluated, recipient age and donor age were the only features consistently influencing graft failure throughout these eras, while features such as glomerular filtration rate and recipient ethnicity showed high importance in specific eras, resulting in relatively poor historical transportability of the best model. Our study emphasises the significance of analysing post-kidney transplant outcomes and identifying era-specific factors mitigating graft failure. The proposed framework can serve as a foundation for future research and assist physicians in identifying patients at risk of graft failure.

Geospatial modeling of wildfire susceptibility on a national scale in Montenegro: A comparative evaluation of F-AHP and FR methodologies

Abstract Wildfires pose a significant ecological, environmental, and socioeconomic challenge in southeastern Europe. The preservation of wildlands is not only essential but also a foremost priority for Montenegro, a country recognized as the world’s first ecological state. Consequently, the development of optimal methodologies and models is of paramount importance to enhance fire protection measures. With this objective in mind, this study strives to create a wildfire susceptibility model on a national scale for Montenegro. The study employed seven natural and anthropogenic causative criteria: vegetation type; aspect; slope; elevation; climate classification; distance from road; and population. The modeling process integrates both natural and anthropogenic causal criteria, employing the Fuzzy Analytic Hierarchy Process (F-AHP) and Frequency Ratio (FR) within geoinformatics environment. The outcomes of the F-AHP model reveal that 72.84% of the total area is categorized as having high to very high susceptibility. Conversely, based on the FR model, only 29.07% of the area falls within these susceptibility levels. In terms of validation, the area under curvature values indicates good performance of the F-AHP model. In contrast, the FR model demonstrates poor performance. These novel findings, pertaining to Montenegro at a national scale, offer valuable insights for preemptive wildfire safeguarding efforts. Moreover, the methodologies employed, with necessary modifications, hold potential for application in geographically diverse regions.

DSEception: a noval neural networks architecture for enhancing pneumonia and tuberculosis diagnosis.

Pneumonia and tuberculosis are prevalent pulmonary diseases globally, each demanding specific care measures. However, distinguishing between these two conditions imposes challenges due to the high skill requirements for doctors, the impact of imaging positions and respiratory intensity of patients, and the associated high healthcare costs, emphasizing the imperative need for intelligent and efficient diagnostic methods. This study aims to develop a highly accurate automatic diagnosis and classification method for various lung diseases (Normal, Pneumonia, and Tuberculosis). We propose a hybrid model, which is based on the InceptionV3 architecture, enhanced by introducing Deepwise Separable Convolution after the Inception modules and incorporating the Squeeze-and-Excitation mechanism. This architecture successfully enables the model to extract richer features without significantly increasing the parameter count and computational workload, thereby markedly improving the performance in predicting and classifying lung diseases. To objectively assess the proposed model, external testing and five-fold cross-validation were conducted. Additionally, widely used baseline models in the scholarly community were constructed for comparison. In the external testing phase, the our model achieved an average accuracy (ACC) of 90.48% and an F1-score (F1) of 91.44%, which is an approximate 4% improvement over the best-performing baseline model, ResNet. In the five-fold cross-validation, our model's average ACC and F1 reached 88.27% ± 2.76% and 89.29% ± 2.69%, respectively, demonstrating exceptional predictive performance and stability. The results indicate that our model holds promise for deployment in clinical settings to assist in the diagnosis of lung diseases, potentially reducing misdiagnosis rates and patient losses. Utilizing deep learning for automatic assistance in the diagnosis of pneumonia and tuberculosis holds clinical significance by enhancing diagnostic accuracy, reducing healthcare costs, enabling rapid screening and large-scale detection, and facilitating personalized treatment approaches, thereby contributing to widespread accessibility and improved healthcare services in the future.

Industrial water withdrawal prediction using multi-head attention encoder model

ABSTRACT Industrial water withdrawal prediction is the cornerstone of water resources monitoring and early warning, as well as a key task in implementing rigid constraints on water resources. Although many factors may influence water withdrawal, they are difficult to obtain. Extracting the features in the historical water withdrawal data itself is a more convenient alternative way to achieve accurate prediction. However, due to the lack of significant regularity and periodicity in industrial water data, feature extraction is challenging. In this paper, a multi-head attention encoder (MAEN) model consisting of multi-head attention mechanisms and feed-forward neural networks is proposed to enable more convenient and accurate forecasting. The multi-head attention mechanism allows to focus on different parts of the data simultaneously, detecting complex dependencies, and extracting key features from historical data to enable more efficient feature extraction. Applied to water withdrawal data from real factories, the proposed model outperforms common time series prediction models including artificial neural network, long short-term memory, and gate recurrent unit. Compared to the best-performing comparison model, MAEN shows a reduction of mean squared error by 3.2% for 1 day ahead predictions, and 8.9% for 7 day ahead predictions, illustrating its superior performance in industrial water withdrawal prediction.

Promoting Sustainable Development of Coal Mines: CNN Model Optimization for Identification of Microseismic Signals Induced by Hydraulic Fracturing in Coal Seams

Borehole hydraulic fracturing in coal seams can prevent dynamic coal mine disasters and promote the sustainability of the mining industry, and microseismic signal recognition is a prerequisite and foundation for microseismic monitoring technology that evaluates the effectiveness of hydraulic fracturing. This study constructed ultra-lightweight CNN models specifically designed to identify microseismic waveforms induced by borehole hydraulic fracturing in coal seams, namely Ul-Inception28, Ul-ResNet12, Ul-MobileNet17, and Ul-TripleConv8. The three best-performing models were selected to create both a probability averaging ensemble CNN model and a voting ensemble CNN model. Additionally, an automatic threshold adjustment strategy for CNN identification was introduced. The relationships between feature map entropy, training data volume, and model performance were also analyzed. The results indicated that our in-house models surpassed the performance of the InceptionV3, ResNet50, and MobileNetV3 models from the TensorFlow Keras library. Notably, the voting ensemble CNN model achieved an improvement of at least 0.0452 in the F1 score compared to individual models. The automatic threshold adjustment strategy enhanced the identification threshold’s precision to 26 decimal places. However, a continuous zero-entropy value in the feature maps of various channels was found to detract from the model’s generalization performance. Moreover, the expanded training dataset, derived from thousands of waveforms, proved more compatible with CNN models comprising hundreds of thousands of parameters. The findings of this research significantly contribute to the prevention of dynamic coal mine disasters, potentially reducing casualties, economic losses, and promoting the sustainable progress of the coal mining industry.

Polygenic risk score portability for common diseases across genetically diverse populations

BackgroundPolygenic risk scores (PRS) derived from European individuals have reduced portability across global populations, limiting their clinical implementation at worldwide scale. Here, we investigate the performance of a wide range of PRS models across four ancestry groups (Africans, Europeans, East Asians, and South Asians) for 14 conditions of high-medical interest.MethodsTo select the best-performing model per trait, we first compared PRS performances for publicly available scores, and constructed new models using different methods (LDpred2, PRS-CSx and SNPnet). We used 285 K European individuals from the UK Biobank (UKBB) for training and 18 K, including diverse ancestries, for testing. We then evaluated PRS portability for the best models in Europeans and compared their accuracies with respect to the best PRS per ancestry. Finally, we validated the selected PRS models using an independent set of 8,417 individuals from Biobank of the Americas-Genomelink (BbofA-GL); and performed a PRS-Phewas.ResultsWe confirmed a decay in PRS performances relative to Europeans when the evaluation was conducted using the best-PRS model for Europeans (51.3% for South Asians, 46.6% for East Asians and 39.4% for Africans). We observed an improvement in the PRS performances when specifically selecting ancestry specific PRS models (phenotype variance increase: 1.62 for Africans, 1.40 for South Asians and 0.96 for East Asians). Additionally, when we selected the optimal model conditional on ancestry for CAD, HDL-C and LDL-C, hypertension, hypothyroidism and T2D, PRS performance for studied populations was more comparable to what was observed in Europeans. Finally, we were able to independently validate tested models for Europeans, and conducted a PRS-Phewas, identifying cross-trait interplay between cardiometabolic conditions, and between immune-mediated components.ConclusionOur work comprehensively evaluated PRS accuracy across a wide range of phenotypes, reducing the uncertainty with respect to which PRS model to choose and in which ancestry group. This evaluation has let us identify specific conditions where implementing risk-prioritization strategies could have practical utility across diverse ancestral groups, contributing to democratizing the implementation of PRS.

Effectiveness of South Africa's network of protected areas: Unassessed vascular plants predicted to be threatened using deep neural networks are all located in protected areas.

Globally, we are in the midst of a biodiversity crisis and megadiverse countries become key targets for conservation. South Africa, the only country in the world hosting three biodiversity hotspots within its borders, harbours a tremendous diversity of at-risk species deserving to be protected. However, the lengthy risk assessment process and the lack of required data to complete assessments is a serious limitation to conservation since several species may slide into extinction while awaiting risk assessment. Here, we employed a deep neural network model integrating species climatic and geographic features to predict the conservation status of 116 unassessed plant species. Our analysis involved in total of 1072 plant species and 96,938 occurrence points. The best-performing model exhibits high accuracy, reaching up to 83.6% at the binary classification and 56.8% at the detailed classification. Our best-performing model at the binary classification predicts that 32% (25 species) and 8% (3 species) of Data Deficient and Not-Evaluated species respectively, are likely threatened, amounting to a proportion of 24.1% of unassessed species facing a risk of extinction. Interestingly, all unassessed species predicted to be threatened are in protected areas, revealing the effectiveness of South Africa's network of protected areas in conservation, although these likely threatened species are more abundant outside protected areas. Considering the limitation in assessing only species with available data, there remains a possibility of a higher proportion of unassessed species being imperilled.

Integrating StEP-COMPAC Definition and Enhanced Recovery after Surgery Status in a Machine-learning-based Model for Postoperative Pulmonary Complications in Laparoscopic Hepatectomy

BackgroundPostoperative pulmonary complications (PPCs) contribute to high mortality rates and impose significant financial burdens. In this study, a machine learning-based prediction model was developed to identify patients at high risk of developing PPCs following laparoscopic hepatectomy. MethodsData were collected from 1022 adult patients who underwent laparoscopic hepatectomy at two centres between January 2015 and February 2022. The dataset was divided into a development set and a temporal external validation set based on the year of surgery. A total of 42 factors were extracted for pre-modelling, including the implementation status of Enhanced Recovery after Surgery (ERAS). Feature selection was performed using the least absolute shrinkage and selection operator (LASSO) method. Model performance was assessed using the area under the receiver operating characteristic curve (AUC). The model with the best performance was externally validated using temporal data. ResultsThe incidence of PPCs was 8.7%. Lambda.1se was selected as the optimal lambda for LASSO feature selection. For implementation of ERAS, serum gamma-glutamyl transferase levels, malignant tumour presence, total bilirubin levels, and age-adjusted Charleston Comorbidities Index were the selected factors. Seven models were developed. Among them, logistic regression demonstrated the best performance, with an AUC of 0.745 in the internal validation set and 0.680 in the temporal external validation set. ConclusionsBased on the most recent definition, a machine learning model was employed to predict the risk of PPCs following laparoscopic hepatectomy. Logistic regression was identified as the best-performing model. ERAS implementation was associated with a reduction in the number of PPCs.

Data-Driven Cycle Life Prediction of Lithium Metal-Based Rechargeable Battery Based on Discharge/Charge Capacity and Relaxation Features.

Achieving precise estimates of battery cycle life is a formidable challenge due to the nonlinear nature of battery degradation. This study explores an approach using machine learning (ML) methods to predict the cycle life of lithium-metal-based rechargeable batteries with high mass loading LiNi0.8Mn0.1Co0.1O2 electrode, which exhibits more complicated and electrochemical profile during battery operating conditions than typically studied LiFePO₄/graphite based rechargeable batteries. Extracting diverse features from discharge, charge, and relaxation processes, the intricacies of cell behavior without relying on specific degradation mechanisms are navigated. The best-performing ML model, after feature selection, achieves an R2 of 0.89, showcasing the application of ML in accurately forecasting cycle life. Feature importance analysis unveils the logarithm of the minimum value of discharge capacity difference between 100 and 10 cycle (Log(|min(ΔDQ 100-10(V))|)) as the most important feature. Despite the inherent challenges, this model demonstrates a remarkable 6.6% test error on unseen data, underscoring its robustness and potential for transformative advancements in battery management systems. This study contributes to the successful application of ML in the realm of cycle life prediction for lithium-metal-based rechargeable batteries with practically high energy density design.

Interpretable Machine Learning Model Based on Superb Microvascular Imaging for Non-Invasive Determination of Crescent Status of IgAN.

To assess the crescentic status of IgA nephropathy (IgAN) non-invasively using a superb microvascular imaging (SMI)-based radiomics machine learning (ML) model. IgAN patients who underwent renal biopsy from June 2022 to October 2023, with two-dimensional ultrasound (US) and SMI examinations conducted one day prior to the renal biopsy. The patients selected were divided randomly into a training group and a test group in a 7:3 ratio. Radiomic features were extracted from US and SMI images, then radiomic features were constructed and ML models were further established using logistic regression (LR) and extreme gradient boosting (XGBoost)XGBoost to determine the crescentic status. The utility of the proposed model was evaluated using receiver operating characteristics, calibration, and decision curve analysis. The SHapley Additive exPlanations (SHAP) was utilized to explain the best-performing ML model. A total of 147 IgAN patients were included in the study, with 103 in the training group and 44 in the test group .Among them, the US-SMI based XGBoost model achieved the best results, with an the area under the curve (AUC) of 0.839 (95% CI,0.756-0.910) and an accuracy of 78.6% in the training group.In the test group, the AUC was 0.859 (95% CI,0.721-0.964), and the accuracy was 81.8%, significantly surpassing the ML model of a single modality and the clinical model established based on occult blood. Additionally, the decision curve analysis (DCA) demonstrated that the XGBoost model provided a higher overall net benefit in the both groups. The SMI radiomics ML model has the capability to accurately predict the crescentic status of IgAN patients, providing effective assistance for clinical treatment decisions.