ML Models Research Articles

Well-Tempered Medical Prompt Engineering for Explainable Extubation.

This study investigated whether the large language model (LLM) utilizes sufficient domain knowledge to reason about critical medical events such as extubation. In detail, we tested whether the LLM accurately comprehends given tabular data and variable importance and whether it can be used in complement to existing ML models such as XGBoost.

A Novel Method and Python Library for ECG Signal Quality Assessment.

Electrocardiogram (ECG) is one of the reference cardiovascular diagnostic exams. However, the ECG signal is very prone to being distorted through different sources of artifacts that can later interfere with the diagnostic. For this reason, signal quality assessment (SQA) methods that identify corrupted signals are critical to improve the robustness of automatic ECG diagnostic methods. This work presents a review and open-source implementation of different available indices for SQA as well as introducing an index that considers the ECG as a dynamical system. These indices are then used to develop machine learning models which evaluate the quality of the signals. The proposed index along the designed ML models are shown to improve SQA for ECG signals.

Multi-Objective Performance Optimization of Machine Learning Models in Healthcare.

Multi-objective optimization holds particular significance for medical applications, wherein enhancing sensitivity is crucial to avoid costly missed diagnoses, and maintaining high specificity is imperative to prevent unnecessary procedures. In particular, when optimizing machine learning architectures for clinical diagnostics, it becomes essential to balance target quality measures such as accuracy, sensitivity, and specificity. Therefore, we developed MOOF, a multi-objective optimization framework that employs NSGA-II and TOPSIS to simultaneously optimize the model parameters of three selected ML algorithms: random forest, support vector machine, and multilayer perceptron. Finally, we evaluated the performance of the optimized MOOF models compared to gold standard methods such as multi-score grid search and single objective optimizations. Our results show that MOOF generally outperforms other approaches by inherently providing optimal solutions, representing the trade-offs between the target objectives. In conclusion, the study supports the importance of multi-objective optimization in medical informatics, with MOOF as a powerful tool for precise ML models, potentially improving patient care and clinical decision support systems.

Swin-chart: An efficient approach for chart classification

Charts are a visualization tool used in scientific documents to facilitate easy comprehension of complex relationships underlying data and experiments. Researchers use various chart types to convey scientific information, so the problem of data extraction and subsequent chart understanding becomes very challenging. Many studies have been taken up in the literature to address the problem of chart mining, whose motivation is to facilitate the editing of existing charts, carry out extrapolative studies, and provide a deeper understanding of the underlying data. The first step towards chart understanding is chart classification, for which traditional ML and CNN-based deep learning models have been used in the literature. In this paper, we propose Swin-Chart, a Swin transformer-based deep learning approach for chart classification, which generalizes well across multiple datasets with a wide range of chart categories. Swin-Chart comprises a pre-trained Swin Transformer, a finetuning component, and a weight averaging component. The proposed approach is tested on a five-chart image benchmark dataset. We observed that the Swin-Chart model outperformers existing state-of-the-art models on all the datasets. Furthermore, we also provide an ablation study of the Swin-Chart model with all five datasets to understand the importance of various sub-parts such as the back-bone Swin transformer model, the value of several best weights selected for the weight averaging component, and the presence of the weight averaging component itself.The Swin-Chart model also received first position in the chart classification task on the latest dataset in the CHART Infographics competition at ICDAR 2023 - chartinfo.github.io.

Modelling and predicting lift force and trans-membrane pressure using linear, KNN, ANN and response surface models during the separation of oil drops from produced water

This study aims to improve the efficiency and reliability of membrane filtration systems used to treat produced water by accurately predicting lift force, transmembrane pressure (TMP), and total resistance (TR) using ML models, namely linear, K-nearest neighbors (KNN), and artificial neural networks (ANN). The dataset was split into training, validation, and test sets, allowing for comprehensive training and evaluation of model performance. Linear, KNN and ANN models achieved high R-squared scores (R2) of 0.998, 0.996, and 0.999 and low mean squared error (MSE) values of 0.0000046, 0.0002763, and 0.000004621 for Lift force prediction, respectively. For TMP prediction, these models showed R2 values of 0.801, 0.90, and 0.72 and MSE values of 0.01236, 0.109, and 0.283, respectively. Similarly, these models predicted TR with R2 scores of 0.76, 0.78, and 0.78 and MSE values of 0.2087, 0.025, and 0.22, respectively. These results indicate that KNN is particularly efficient in accurately predicting the membrane parameters and could be a robust tool to optimize membrane filtration processes. Response surface modelling (RSM) elucidates the relationship between input and output parameters. Initially, the increase in drop size and shear has a minimal impact on the lift force. Beyond a drop size of 0.3 μm and shear rate 1200 S−1, lift force rises significantly. TMP initially increases with shear rate, peaking at a value of 1200 S−1, then decreases with higher shear rates. Optimal conditions identified are higher shear rates and drop sizes >0.3 μm to maintain lower TMP and effective filtration. Findings from the study suggest that employing predictive modelling can significantly enhance the efficiency and reliability of produced water treatment, by improving contaminant removal, reducing energy consumption and cost, potentially reducing environmental impacts, and improving sustainability in the oil and gas industry.

Identification of avoidable patients at triage in a Paediatric Emergency Department: a decision support system using predictive analytics

BackgroundCrowding has been a longstanding issue in emergency departments. To address this, a fast-track system for avoidable patients is being implemented in the Paediatric Emergency Department where our study is conducted. Our goal is to develop an optimized Decision Support System that helps in directing patients to this fast track. We evaluated various Machine Learning models, focusing on a balance between complexity, predictive performance, and interpretability.MethodsThis is a retrospective study considering all visits to a university-affiliated metropolitan hospital’s PED between 2014 and 2019. Using information available at the time of triage, we trained several models to predict whether a visit is avoidable and should be directed to a fast-track area.ResultsA total of 507,708 visits to the PED were used in the training and testing of the models. Regarding the outcome, 41.6% of the visits were considered avoidable. Except for the classification made by triage rules, i.e. considering levels 1,2, and 3 as non-avoidable and 4 and 5 as avoidable, all models had similar results in model’s evaluation metrics, e.g. Area Under the Curve ranging from 74% to 80%.ConclusionsRegarding predictive performance, the pruned decision tree had evaluation metrics results that were comparable to the other ML models. Furthermore, it offers a low complexity and easy to implement solution. When considering interpretability, a paramount requisite in healthcare since it relates to the trustworthiness and transparency of the system, the pruned decision tree excels.Overall, this paper contributes to the growing body of research on the use of machine learning in healthcare. It highlights practical benefits for patients and healthcare systems of the use ML-based DSS in emergency medicine. Moreover, the obtained results can potentially help to design patients’ flow management strategies in PED settings, which has been sought as a solution for addressing the long-standing problem of overcrowding.

Air Quality Analysis through IoT Device and Risk Prediction of Asthma Attack using ML Techniques

Background: The air quality of any area depends upon the various PMs (particulate matter) and hazardous gases present in the air. Low-cost PM sensors and gas sensors are present in different target places to monitor the air quality, read the environmental data, and transmit it to local servers through the IoT device. The low-cost sensor is not reliable due to its low sensing capacity; therefore, the read data is calibrated with the meteorological data presented by the nearby meteorological Centre of that particular area. The calibrated reading data sent to the server could be analyzed through some Machine Learning [ML] models. The ML models help to predict the risk of asthma in a particular area. The risk of asthma is directly related to the air quality of the surroundings. It is observed that the air quality of the industrial area is much worse than the non-industrial belt. Air quality monitoring of industrial areas is always a challenging task due to the ununiformed pollution in some segregated places around the industry, emitting pollutants mostly from chimneys. The air quality of any area depends upon the PM (PM), i.e., PM2.5 and PM10.0, as well as the gasses like NO2(Nitrogen Dioxide), NH3 (Ammonia), SO2(Sulfur dioxide), CO(Carbon monoxide), O3(Ozone) and Benzene. These are the most hazardous gases generally emitted by common heavy industries like iron and steel. In this article, the researchers considered the industrial belt of the Asansol- Durgapur region of West Bengal, India, and predicted the risk of asthma attacks for the test dataset. The experiment was carried out on 10 different supervised machine learning [SML] models as well as semi-supervised machine learning (SSML) models. The SML models have been further refined through hyper-parameter tuning, and better results have been obtained in the case of some ML models. The result has been compared with the existing literature considering the same external environment from where the meteorological data was collected, and similar ML models have been used. The research outperformed the existing literature, which is depicted in the result and analysis section of the article. Methods: The study evaluated ML models, both supervised and semi-supervised, to assess pollution levels. Relevant features were selected while less relevant ones were discarded. Accuracy levels of different ML algorithms werecompared in the results. The most effective model for an IoT system was chosen to maximize accuracy. In semi-supervised learning, feature selection followed supervised learning, but testing was akin to unsupervised learning. Results were compared with supervised learning data, enhancing reliability. Results: The result employing various classifiers werepresented across tables after the independent parameter Ozone was removed. Following the output of several classifiers, the results were verified using the k-fold validation method, with k being set to 5 or 10, accordingly. Tables display the best outcome, which is indicated in bold characters. method: In this research work the researcher considered 9 different ML models and used them as supervised as well as semi supervised model to determine the pollution level of the certain area. In this research work the researcher also selected the most relevant features and discarded the less relevant features. In case of SML algorithm, the accuracy level of the different ML algorithm has been determined and depicted in the result analysis section. The most effective ML model has been chosen for the proposed embedded system so that accuracy could be achieved at most. In case of semi supervised algorithm the feature selection is done as per the supervised algorithm. In this case the training is done same as the SML algorithm, but the testing phase is done like unsupervised machine learning algorithm where the decision parameter is predicted and ultimately matched with the previously achieved data of SML algorithm. The reliability of this approach is much more effective than simple SML algorithm. Conclusion: This study focused on predicting asthma risk in the Asansol-Durgapur industrial belt, India, using low-cost PM and gas sensors. Data calibration with meteorological inputs enhanced accuracy. ML models predicted risk and were refined through hyper-parameter tuning. Comparative analysis showed superior performance, emphasizing the importance of precise air quality monitoring. While offering a robust framework for future research, the study’s limitation lies in its area-specific dataset.

Early sepsis mortality prediction model based on interpretable machine learning approach: development and validation study.

Sepsis triggers a harmful immune response due to infection, causing high mortality. Predicting sepsis outcomes early is vital. Despite machine learning's (ML) use in medical research, local validation within the Medical Information Mart for Intensive Care IV (MIMIC-IV) database is lacking. We aimed to devise a prognostic model, leveraging MIMIC-IV data, to predict sepsis mortality and validate it in a Chinese teaching hospital. MIMIC-IV provided patient data, split into training and internal validation sets. Four ML models logistic regression (LR), support vector machine (SVM), deep neural networks (DNN), and extreme gradient boosting (XGBoost) were employed. Shapley additive interpretation offered early and interpretable mortality predictions. Area under the ROC curve (AUROC) gaged predictive performance. Results were cross verified in a Chinese teaching hospital. The study included 27,134 sepsis patients from MIMIC-IV and 487 from China. After comparing, 52 clinical indicators were selected for ML model development. All models exhibited excellent discriminative ability. XGBoost surpassed others, with AUROC of 0.873 internally and 0.844 externally. XGBoost outperformed other ML models (LR: 0.829; SVM: 0.830; DNN: 0.837) and clinical scores (Simplified Acute Physiology Score II: 0.728; Sequential Organ Failure Assessment: 0.728; Oxford Acute Severity of Illness Score: 0.738; Glasgow Coma Scale: 0.691). XGBoost's hospital mortality prediction achieved AUROC 0.873, sensitivity 0.818, accuracy 0.777, specificity 0.768, and F1 score 0.551. We crafted an interpretable model for sepsis death risk prediction. ML algorithms surpassed traditional scores for sepsis mortality forecast. Validation in a Chinese teaching hospital echoed these findings.

An interpretable machine learning methodology to generate interaction effect hypotheses from complex datasets

AbstractMachine learning (ML) models are increasingly being used in decision‐making, but they can be difficult to understand because most ML models are black boxes, meaning that their inner workings are not transparent. This can make interpreting the results of ML models and understanding the underlying data‐generation process (DGP) challenging. In this article, we propose a novel methodology called Simple Interaction Finding Technique (SIFT) that can help make ML models more interpretable. SIFT is a data‐ and model‐agnostic approach that can be used to identify interaction effects between variables in a dataset. This can help improve our understanding of the DGP and make ML models more transparent and explainable to a wider audience. We test the proposed methodology against various factors (such as ML model complexity, dataset noise, spurious variables, and variable distributions) to assess its effectiveness and weaknesses. We show that the methodology is robust against many potential problems in the underlying dataset as well as ML algorithms.

Optimization and prediction of dye adsorption utilising cross-linked chitosan-activated charcoal: Response Surface Methodology and machine learning

Water pollution poses a significant environmental threat due to the discharge of organic dyes from industrial processes. In this study, we investigated a novel adsorptive composite material for removing MB dye from water and applied new models for optimizing and predicting adsorption efficiency. The cross-linked chitosan-activated charcoal composite material was synthesized using a single-step process and characterized by analytical techniques including FTIR, XPS, BET, SEM, TGA, and zeta potential, which demonstrated its remarkable physicochemical properties. Batch adsorption experiments showed a direct correlation between increased adsorbent dosage and enhanced MB dye adsorption, with an optimal dosage of 0.4 g/25 mL. Additionally, pH-dependent studies revealed excellent performance at lower pH levels, achieving up to 95 % adsorption. The composite reached equilibrium within the first 360 min of contact time and maintained a stable 99 % adsorption efficiency afterward. These results were confirmed by an applied optimization RSM model. Additionally, adsorption isotherms and kinetics were thoroughly analyzed by fitting them into various models, revealing a maximum dye adsorption capacity of 625 mg/g. ML models, such as RF, XGBoost and stacking, were used to predict adsorption efficiency, with a stacked model combining RF and XGBoost showing the highest predictive accuracy. Variable importance plots indicated that contact time was the most significant factor influencing adsorption, followed by dosage and pH, in line with RSM findings. This research highlights the eco-friendliness and cost-effectiveness of the composite material in addressing water pollution challenges. Integrating ML models and RSM enhances the predictability and optimization of the adsorption process, offering valuable insights for mitigating the impact of organic dyes on aquatic ecosystems and public health.

Geotechnical and microstructural analysis of high-volume fly ash stabilized clayey soil and machine learning application

The weak soil stabilization using solid wastes is one of the most common solutions for improving geotechnical characteristics as well as for problematic waste dumping in landfills. The present experimental study aims to examine the effect of high-volume Class-F fly ash on the geotechnical and microstructural properties of clayey soil by adding them in ranges between 5 % and 50 %. The results show that as the amount of fly ash in clayey soil increases, properties like the specific gravity, plasticity index, permeability, optimum moisture content, maximum dry density and free swelling index improves. Moreover, these geotechnical properties were analyzed to develop machine learning models using three different algorithms, namely K-nearest neighbor regression, random forest, and support vector regression, for obtaining the optimum amount of fly ash contents in weak expansive soils. The predicted and experimental results found to be in close-relation for predicting the geotechnical behavior of modified clayey soil. Furthermore, the performance of the ML models degrades as the number of components reduces, with KNN regression consistently outperforming SVR and RF but suffering significantly with fewer components. The results of the testing set in the case of four components are MSE of 77, R² of 0.896, RMSE of 0.846, MAE of 0.327, and SEE of 0.858, indicating precise and consistent predictions. However, the prediction accuracy considering lesser components shows MSE as 262, R² as 0.648, MAE as 5.606, SEE as 16.707, and GPI as 1.056, confirming the elevated error rates. Overall, it has been concluded that combining comprehensive experimental work and machine learning techniques outperforms in enhancing geotechnical data processing, optimized waste contents in weak soils, improves sustainability in construction, saves resources, reduces the possibility of human mistakes, and increases reliability.

A comparative study of predicting high entropy alloy phase fractions with traditional machine learning and deep neural networks

Predicting phase stability in high entropy alloys (HEAs), such as phase fractions as functions of composition and temperature, is essential for understanding alloy properties and screening desirable materials. Traditional methods like CALPHAD are computationally intensive for exploring high-dimensional compositional spaces. To address such a challenge, this study explored and compared the effectiveness of random forests (RF) and deep neural networks (DNN) for accelerating materials discovery by building surrogate models of phase stability prediction. For interpolation scenarios (testing on the same order of system as trained), RF models generally produce smaller errors than DNN models. However, for extrapolation scenarios (training on lower-order systems and testing on higher order systems), DNNs generalize more effectively than traditional ML models. DNN demonstrate the potential to predict topologically relevant phase composition when data were missing, making it a powerful predictive tool in materials discovery frameworks. The study uses a CALPHAD dataset of 480 million data points generated from a custom database, available for further model development and benchmarking. Experiments show that DNN models are data-efficient, achieving similar performance with a fraction of the dataset. This work highlights the potential of DNNs in materials discovery, providing a powerful tool for predicting phase stability in HEAs, particularly within the Cr-Hf-Mo-Nb-Ta-Ti-V-W-Zr composition space.

Enhancing analyst decisions for seismic source discrimination with an optimized learning model

Sustainable development in urban areas requires a wide variety of current and theme-based information for efficient management and planning. In addition, researching the spatial distribution of earthquake (EQ) clusters is an important step in reducing seismic risks and EQ losses through better assessment of seismic hazards, therefore it is desirable to acquire an uncontaminated database of seismic activity. Quarry blasts (QBs) conducted over the mapped area have tainted the seismicity inventory in the northwestern region of Egypt, which is the focus of this paper. Separating these QBs from the EQs is hence preferable for accurate seismicity and risk assessments. Consequently, we present a highly effective ML model for cleaning up the seismicity database, allowing for the accurate delineation of EQ clusters using data from a single seismic station, “AYT”, which is part of the Egyptian National Seismic Network. The magnitudes ≤3\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\le 3$$\\end{document} that are very uncertain as EQs or QBs and need a significant amount of time to analyze are the primary focus of the model. In order to find the best way to classify EQs and QBs, the method looks at a number of ML models before settling on the best one using eight features. The results show that the suggested method, which uses the quadratic discrimination analysis model for discriminating, successfully separates EQs and QBs with a 99.4% success rate.

The application of transfer machine learning to predict and impute missing sulphate levels in different Acid Mine Drainage treatment plants

An accurately pre-trained stacking ensemble machine learning regressor was used to predict sulphate levels in two other Acid Mine Drainage (AMD) treatment plants using Transfer Learning (TL). The model was trained on the large Central Rand (CR) water quality dataset and was used to predict and impute the sulphate levels in the scanty East Rand (ER) and West Rand (WR) datasets which would not have been sufficient to train ML models from scratch. TL was successfully used to overcome this barrier and rapidly predicted sulphate levels in the East Rand and West Rand plants using the pre-trained model and achieved a high level of accuracy (Mean Squared Error:0.00124, Mean Absolute Error:0.0290 and R2:0.963) for the East Rand plant when comparing the predicted and true sulphate values. No true sulphate values existed for the West Rand plant; however, TL was successful in imputing these missing values and rapidly completed the West Rand dataset by providing the historic sulphate levels. This was possible due to the high degree of similarity between all domains (treatment plants) since they had similar geographic locations, the same treatment process, possessed the same important features and had the same relationships between variables. TL was successful in providing three accurate datasets for AMD sulphate levels, an important accomplishment towards having reliable data for use in design of experiments aimed at recovering valuable resources such as elemental sulphur, gypsum and important metals from AMD.

Predicting Particle Size and Soil Organic Carbon of Soil Profiles Using VIS-NIR-SWIR Hyperspectral Imaging and Machine Learning Models

Modeling spectral reflectance data using machine learning algorithms presents a promising approach for estimating soil attributes. Nevertheless, a comprehensive investigation of the most effective models, parameters, wavelengths, and data acquisition techniques is essential to ensure optimal predictive accuracy. This work aimed to (a) explore the potential of the soil spectral signature obtained in different spectral bands (VIS-NIR, SWIR, and VIS-NIR-SWIR) and, by using hyperspectral imaging and non-imaging sensors, in the predictive modeling of soil attributes; and (b) analyze the accuracy of different ML models in predicting particle size and soil organic carbon (SOC) applied to the spectral signature of different spectral bands. Six soil monoliths, located in the central north region of Parana, Brazil, were collected and scanned via hyperspectral cameras (VIS-NIR camera and SWIR camera) and spectroradiometer (VIS-NIR-SWIR) in the laboratory. The spectral signature of the soils was analyzed and subsequently applied to ML models to predict particle size and SOC. Each set of data obtained by the different sensors was evaluated separately. The algorithms used were k-nearest neighbors (KNN), support vector machine (SVM), random forest (RF), linear regression (LR), artificial neural network (NN), and partial least square regression (PLSR). The most promising predictive performance was observed for the complete VIS-NIR-SWIR spectrum, followed by SWIR and VIS-NIR. Meanwhile, KNN, RF, and NN models were the most promising algorithms in estimating soil attributes for the dataset obtained from both sensors. The general mean R2 (determination coefficient) values obtained using these models, considering the different spectral bands evaluated, were around 0.99, 0.98, and 0.97 for sand prediction, and around 0.99, 0.98, and 0.96 for clay prediction. The lower performances, obtained for the datasets from both sensors, were observed for silt and SOC, with R2 results between 0.40 and 0.59 for these models. KNN demonstrated the best predictive performance. Integrating effective ML models with robust sample databases, obtained by advanced hyperspectral imaging and spectroradiometers, can enhance the accuracy and efficiency of soil attribute prediction.

Development of laminar burning velocity prediction model and correlation of iso-octane air mixtures using artificial neural network

Surrogate fuels provide an economical alternative for forecasting the combustion properties of transportation fuels like diesel, gasoline, and kerosene. Iso-octane (2,2,4-trimethylpentane), a primary reference fuel for gasoline, is extensively used as a surrogate component. This study utilizes a Feed-Forward Artificial Neural Network (FFANN) with back-propagation (BP) to predict the laminar burning velocity (LBV) of iso-octane/air mixtures. The ANN model was developed using a dataset of 6339 data points, including 3788 experimental data points from literature since 2004 and 2551 simulation data points from reduced kinetic reaction mechanism (RKM) CHEMKIN simulations. The grid search cross-validation (CV) method was employed to optimize the ANN hyperparameters. Implemented in Python using Keras, the ANN model addresses research gaps by forecasting LBV for hydrocarbons beyond n-heptane and modelling LBV under conditions reflective of actual engine operations. Unlike prior studies, we optimized various ML models, their complexity, size, and hyperparameters to achieve high prediction accuracy. We also integrated a hybrid model combining Particle Swarm Optimization (PSO) with Genetic Algorithms (GA) for hyperparameter optimization, ensuring efficient configuration of the ANN. The constructed ANN model was compared to other ML models, including generalized linear regression (GLR), support vector machine (SVM), random forest (RF), and XGBoost regression. For the testing set (20 % of the total dataset), the ANN model outperformed all other ML models and empirical correlations, achieving a coefficient of determination (R2) of 0.9903, root mean square error (RMSE) of 1.911, and mean absolute error (MAE) of 1.206. Optimizing the ANN model with PSO further enhanced prediction accuracy, achieving a correlation coefficient of 0.9973 and a reduced mean absolute error of 0.753. To evaluate the ANN model's predictive ability, a reduced mechanism from the Lawrence Livermore National Laboratory (LLNL) was used. Compared to the LLNL RKM, the ANN predicted values showed lower percentage deviation from experimental data. Additionally, computation cost comparisons revealed that while RKM simulations in CHEMKIN took 244.2 s per case, the PSO-optimized ANN model required only 900 s for 150 cases. A 16-term correlation for laminar burning velocity (LBV) was derived from ANN predictions, serving as a practical tool to predict LBV based on pressure, temperature, and equivalence ratio. Unlike existing literature correlations, this newly developed correlation is valid across a wide range of operating conditions, offering a fundamental database for real-time combustion applications.

Design of a precise ensemble expert system for crop yield prediction using machine learning analytics

AbstractAgriculture is facing significant challenges in the development of crop yield forecasts, which are important aspects of decision‐making at the international, regional, and local levels. The area of agriculture is attracting growing attention because of increasing the demand for food supplies. To ensure future food supplies, crop yield prediction (CYP) provides the best decision‐making to assist farmers in agricultural yield forecasting efficiently. Nevertheless, CYP is a difficult endeavor because of the intricacy of the underlying mechanisms and the effect of numerous factors, including weather patterns, soil characteristics, and crop management techniques. In today's era, ensemble learning (EL) approaches have recently demonstrated significant promise for enhancing the reliability and accuracy of CYP. The success of the EL techniques depends on several facts, including how the base learner models are trained and how these are combined. This study provides important insights into the EL techniques for CYP. This paper proposes an expert system model named precise ensemble expert system for crop yield prediction (PEESCYP) to predict the best crop for agricultural land. The proposed PEESCYP model employs multiple imputation by chained equation (MICE) data imputation technique to treat the missing values of the collected dataset, the isolation forest (IF) technique for outlier detection, the ant colony optimization (ACO) technique to perform feature selection, robust scaling (RS) technique to perform data normalization, and the extra tree (ET) is used for classification to overcome the variance and overfitting problem of the single classifiers. The measurements of the proposed PEESCYP model have been collected by means of accuracy, precision, recall, and F‐1 score using a prepared dataset, which is collected from International Crops Research Institute for the Semi‐Arid Tropics (ICRISAT), and the proposed model is compared with different single‐classifier based ML models, EL models, and various existing models available in the literature. The results of this experiment underline that the proposed PEESCYP model outperforms the others.

Combining global precipitation data and machine learning to predict flood peaks in ungauged areas with similar climate

Increasing flood risk due to urbanization and climate change poses a significant challenge to societies at global scale. Hydrologic information that is required for understanding flood processes and for developing effective warning procedures is currently lacking in most parts of the world. Procedures that can combine global climate dataset from satellite and reanalysis with fast and low computational cost prediction systems, are attractive solutions for addressing flood predictions in ungauged areas. This work develops and tests a prediction framework that relies on two fundamental components. First, meteorological data from global datasets (IMERG and ERA5-Land) provide key input variables and second, ML models trained in the data-rich contiguous US, are applied in climatically similar regions in other parts of the world. Catchments in Australia, Brazil, Chile, Switzerland, and Great Britain were used as pseudo-ungauged regions for testing. Results indicate acceptable performance for both IMERG and ERA5-Land forced models with relative difference in flood peak prediction within 30 % and similar overall performance to locally trained ML models. Specific climate regions for which ML models have revealed good performance include Mediterranean climates like the US West Coast, subtropical areas like the Southern Atlantic Gulf, and mild temperate regions like the Mid-Atlantic Basin. This work highlights the potential of combining global precipitation dataset with pre-trained ML models in data-rich areas, for flood prediction in ungauged areas with similar climate.

Machine Learning-based Nomograms for Predicting Clinical Stages of Initial Prostate Cancer: A Multicenter Retrospective Study

ObjectivesTo construct and externally validate machine learning-based nomograms for predicting progression stages of initial prostate cancer (PCa) using biomarkers and clinicopathologic features. Methods362 inpatients diagnosed with PCa at the First Affiliated Hospital were randomly assigned to training and testing sets in a 3:7 ratio, while 136 PCa patients from People's Hospital formed the external validation set. Imaging and clinicopathologic information were collected. Optimal features distinguishing advanced prostate cancer (APC) and metastatic PCa (mPCa) were identified through logistic regression (LR). ML algorithms were employed to build and compare ML models. The best-performing algorithm established models for PCa progression stage. Models performance was evaluated using metrics, ROC curves, calibration, and decision curve analysis (DCA) in training, testing, and external validation sets. ResultsFollowing LR analyses, PSA (P=0.001), maximum tumor diameter (P=0.026), Gleason score (P<0.001), and RNF41 (P<0.001) were optimal features for predicting APC, while ALP (P<0.001), PSA (P<0.001), and GS score (P=0.024) were for mPCa. Among ML models, the LR models exhibited superior performance. Consequently, the LR algorithm was used for the APC-risk-nomogram and mPCa-risk-nomogram construction, with AUC values of 0.848, 0.814, 0.810, and 0.940, 0.913, 0.910, in the training, testing, and external validation sets, respectively. Calibration and DCA curves affirmed nomograms' consistency and net benefits for clinical decision-making. ConclusionsIn summary, ML-based APC-risk-nomogram and mPCa-risk-nomogram exhibit outstanding predictive performance for PCa progression stages. These nomograms can assist clinicians in finely categorizing newly diagnosed PCa patients, facilitating personalized treatment plans and prognosis assessment.

Representative Sample Size for Estimating Saturated Hydraulic Conductivity via Machine Learning: A Proof‐Of‐Concept Study

AbstractMachine learning (ML) has been extensively applied in various disciplines. However, not much attention has been paid to data heterogeneity in databases and number of samples used to train ML models in hydrology. In this study, we addressed these issues and their impacts on the accuracy and reliability of ML models in the estimation of saturated hydraulic conductivity, Ks. We selected 17,990 soil samples from the USKSAT database and created random subsets N = 2,000, 4,000, 6,000, 8,000, 10,000, 12,000, 14,000, 16,000, and 17,990, 80% of which were used for training. The random subset selection was repeated 50 times. The extreme gradient boosting (XGBoost) algorithm was used to estimate Ks from other soil properties, such as bulk density, soil depth, texture, and organic content. For each subset, we conducted the learning curve analysis on the training and cross‐validation data sets. Results showed that for all training sample sizes the number of samples was not enough for the training and cross‐validation curves to reach a plateau. We also applied the concept of representative elementary volume by plotting the average coefficient of determination, R2, and root mean square log‐transformed error, RMSLE, against the training sample size. For the testing data set, as the number of training sample size increased from 1,600 to 14,392 the average R2 value increased from 0.74 to 0.90, while the average RMSLE value decreased from 1.08 to 0.69. Either the learning curve or representative sample size analysis is required to investigate whether the number of samples is enough or not.