Search

Hyperparameters play a critical role in analyzing predictive performance in machine learning models. They serve to strike a balance between overfitting and underfitting of research-independent features to prevent extremes. Manual tuning and automated techniques are employed to identify the optimal combination and permutation to achieve the best model performance. This study explores the pursuit of the best fit through various hyperparameters. Following Logistic Regression analysis, this research compared Random Forest, Randomized search, Grid search, Genetic, Bayesian, and Optuna machine learning model tuning for the best accuracy of prediction the student The model accuracy was further assessed using confusion matrices and Receiver Operating Characteristic—Area Under the Curve (ROC-AUC) curves for student grade classification. The genetic algorithm's recommended hyperparameter tuning yielded the highest accuracy (82.5%) and AUC-ROC score (90%) for student result classification. Manual tuning with an estimator of 300, criterion entropy, max features of sqrt, and a minimum sample leaf of 10 achieved an accuracy of 81.1%, which closely resembled the performance randomized search cross-validation algorithm. The default random forest model scored the least accuracy (78%). However, this manual tuning process took a lesser time (3.66 s) to fit the model while grid search CV tuned 941.5 s. Hence, this research made significant contributions to optimizing various machine learning models using a range of hyperparameters for grade classification. [ABSTRACT FROM AUTHOR]

This article investigates the importance of moisture content in cement-stabilized earth blocks (CSEBs) and explores methods for their prediction using machine learning. A key aspect of the research is the development of accurate moisture content prediction models. The study compares the performance of various machine learning models, and XGBoost emerges as the most promising model, demonstrating superior accuracy in predicting moisture content based on factors like soil properties, cement content, and ultrasonic pulse velocity (UPV). The study employs SHAP (SHapley Additive exPlanations) to understand how these features influence the model's predictions. UPV is the most significant factor affecting predicted moisture content, followed by cement content and soil properties like uniformity coefficient. Also, the study explores the possibility of using a reduced set of features for moisture content prediction. They demonstrate that a combination of UPV, cement content, and uniformity coefficient can achieve good accuracy, highlighting the potential for practical applications where obtaining all data points might be challenging. [ABSTRACT FROM AUTHOR]

ABSTRACT Objective Design Setting Participants Main Outcome Measures Results Conclusion Machine learning has been effective in other areas of medicine, this study aims to investigate this with regards to HNC and identify which algorithm works best to classify malignant patients.An observational cohort study.Queen Elizabeth University Hospital.Patients who were referred via the USOC pathway between January 2019 and May 2021.Predicting the diagnosis of patients from three categories, benign, potential malignant and malignant, using demographics and symptoms data.The classic statistical method of ordinal logistic regression worked best on the data, achieving an AUC of 0.6697 and balanced accuracy of 0.641. The demographic features describing recreational drug use history and living situation were the most important variables alongside the red flag symptom of a neck lump.Further studies should aim to collect larger samples of malignant and pre‐malignant patients to improve the class imbalance and increase the performance of the machine learning models. [ABSTRACT FROM AUTHOR]

Summary: As a consequence of continuing demand for additional capacity and higher quality data transmission, channel coding as a key component of a digital communication network has progressed from a classical single pair, point‐to‐point information theoretic application to a class of coding techniques with diverse parameters. The heterogeneous requirements of B5G/6G networks technology have made flexibility of code parameters the main desired feature while designing channel codes. Owing to this, channel coding techniques have evolved to support enabling applications that depend on different factors such as latency, reliability, energy efficiency and throughput. We review and discuss the application of new techniques, modifications in the design and construction of modern channel coding schemes, including block, convolutional, rateless and space‐time codes. Further, the error correction performance of mainstream channel decoders for LDPC, polar and turbo codes is compared and analysed for their attainable error rate. The aim of this study is to highlight the adaptability of the discussed coding schemes for the forthcoming cellular network landscape. This article also addresses the implementational challenges of high throughput, low latency and machine type communication scenarios. Moreover, some recent studies exploring the role of machine learning for optimisation of B5G/6G networks are classified and discussed. Concluding, research areas pertaining to the scalability of error control coding for next generation networking are addressed. [ABSTRACT FROM AUTHOR]

ABSTRACT From early in the coronavirus disease 2019 (COVID‐19) pandemic, there was interest in using machine learning methods to predict COVID‐19 infection status based on vocal audio signals, for example, cough recordings. However, early studies had limitations in terms of data collection and of how the performances of the proposed predictive models were assessed. This article describes how these limitations have been overcome in a study carried out by the Turing‐RSS Health Data Laboratory and the UK Health Security Agency. As part of the study, the UK Health Security Agency collected a dataset of acoustic recordings, SARS‐CoV‐2 infection status and extensive study participant meta‐data. This allowed us to rigorously assess state‐of‐the‐art machine learning techniques to predict SARS‐CoV‐2 infection status based on vocal audio signals. The lessons learned from this project should inform future studies on statistical evaluation methods to assess the performance of machine learning techniques for public health tasks. [ABSTRACT FROM AUTHOR]

Class imbalance is one of many problems of customer churn datasets. One of the common problems is class overlap, where the data have a similar instance between classes. The prediction task of customer churn becomes more challenging when there is class overlap in the data training. In this research, we suggested a hybrid method based on tabular GANs, called CTGAN-ENN, to address class overlap and imbalanced data in datasets of customers that churn. We used five different customer churn datasets from an open platform. CTGAN is a tabular GAN-based oversampling to address class imbalance but has a class overlap problem. We combined CTGAN with the ENN under-sampling technique to overcome the class overlap. CTGAN-ENN reduced the number of class overlaps by each feature in all datasets. We investigated how effective CTGAN-ENN is in each machine learning technique. Based on our experiments, CTGAN-ENN achieved satisfactory results in KNN, GBM, XGB and LGB machine learning performance for customer churn predictions. We compared CTGAN-ENN with common over-sampling and hybrid sampling methods, and CTGAN-ENN achieved outperform results compared with other sampling methods and algorithm-level methods with cost-sensitive learning in several machine learning algorithms. We provide a time consumption algorithm between CTGAN and CTGAN-ENN. CTGAN-ENN achieved less time consumption than CTGAN. Our research work provides a new framework to handle customer churn prediction problems with several types of imbalanced datasets and can be useful in real-world data from customer churn prediction. [ABSTRACT FROM AUTHOR]

This study explores the impact of hyperparameter optimization on machine learning models for predicting cardiovascular disease using data from an IoST (Internet of Sensing Things) device. Ten distinct machine learning approaches were implemented and systematically evaluated before and after hyperparameter tuning. Significant improvements were observed across various models, with SVM and Neural Networks consistently showing enhanced performance metrics such as F1-Score, recall, and precision. The study underscores the critical role of tailored hyperparameter tuning in optimizing these models, revealing diverse outcomes among algorithms. Decision Trees and Random Forests exhibited stable performance throughout the evaluation. While enhancing accuracy, hyperparameter optimization also led to increased execution time. Visual representations and comprehensive results support the findings, confirming the hypothesis that optimizing parameters can effectively enhance predictive capabilities in cardiovascular disease. This research contributes to advancing the understanding and application of machine learning in healthcare, particularly in improving predictive accuracy for cardiovascular disease management and intervention strategies. [ABSTRACT FROM AUTHOR]

Accurate prediction of fatal events in car accidents has significant health management implications. This research article explores the application of imbalanced data handling techniques in machine learning to enhance prediction performance. By implementing these techniques on car accident data, health organizations can identify and forecast a fatal event, enabling more efficient and effective allocation of limited health resources. Concurrently, enhancing the performance of machine learning models through imbalanced data handling techniques can impact health management decisions. Our findings highlight the significance of imbalanced data handling techniques in predicting fatality in car accidents, ultimately contributing to improved road safety and better management of health resources. Moreover, the effective use of imbalanced data demonstrates a substantial improvement in the specificity of the prediction. Addressing the impact of machine learning techniques on imbalanced car accident data can significantly improve overall health outcomes. [ABSTRACT FROM AUTHOR]

Globally, the COVID-19 pandemic has aggravated the problem of depression. In the literature on depression detection, predictive capacity of multi-modal biomarkers like speech, facial and language characteristics were combined to improve the performance of Machine Learning (ML) models. The existing techniques have the limitation of not using the feature based weighting on multimodal data items in supervised learning models. The aim of the study is to use correlation based weighting mechanism between the feature and label in supervised learning to build an automatic multimodal depression detection system. We constructed a dataset of audio-visual recordings that consists of depressed and non-depressed subjects, to address this limitation by proposing a novel two-stage multimodal method for classification. The first stage attention weight is found using the Pearson correlation coefficients. Then these attention weights are combined with the multimodal feature vector, to use in the second stage. The resultant feature vector is experimented with three state-of-the-art classifiers i.e. Support Vector Machine (SVM), Decision Tree (DT) and K-Nearest Neighbour (KNN) classifiers, for predictions. Overall performance measures of proposed model are: accuracy 87.3%, precision 83.3%, recall 77%, and f1-score 80% which showed its effectiveness. Also outcomes of this proposed model were compared with the outcomes of existing state-of-the-art methods, such as baseline base classifiers and feature-level fusion method. Further, the proposed model is also evaluated applying to a widely used benchmarking dataset i.e. Distress Analysis Interview Corpus (DAIC) in depression detection. Results showed that combination of heterogeneous classifiers improves performance when classifying depressed and non-depressed subjects. [ABSTRACT FROM AUTHOR]

The rig state plays a crucial role in recognizing the operations carried out by the drilling crew and quantifying Invisible Lost Time (ILT). This lost time, often challenging to assess and report manually in daily reports, results in delays to the scheduled timeline. In this paper, the Naive Bayes algorithm was used to establish a novel rig state. Training data, consisting of a large set of rules, was generated based on drilling experts' recommendations. This dataset was then employed to build a Naive Bayes classifier capable of emulating the cognitive processes of skilled drilling engineers and accurately recognizing the actual drilling operation from surface data. The developed model was used to process high-frequency drilling data collected from three wells, aiming to derive the Key Performance Indicators (KPIs) related to each drilling crew's efficiency and quantify the ILT during the drilling connections. The obtained results revealed that the established rig state excelled in automatically recognizing drilling operations, achieving a high success rate of 99.747%. The findings of this study offer valuable insights for drillers and rig supervisors, enabling real-time visual assessment of efficiency and prompt intervention to reduce ILT. [ABSTRACT FROM AUTHOR]

In this paper, we attempt to study the effectiveness of various simple machine learning methods in the prediction of bank failures. From a raw dataset of 10,938 US banks during the period of 2000–2020, we find that machine learning approaches do not really outperform the benchmark of conventional statistical method, logistic regression. However, using PCA to retain relevant variance in variables significantly improve the performance of machine learning methods and raise the out-of-sample accuracy of those method to over 70% to over 80%. Of all the machine learning methods used in this paper, the simple KNN seems to be the best model in forecasting bank failure in the United States. [ABSTRACT FROM AUTHOR]

Handling missing data in clinical prognostic studies is an essential yet challenging task. This study aimed to provide a comprehensive assessment of the effectiveness and reliability of different machine learning (ML) imputation methods across various analytical perspectives. Specifically, it focused on three distinct classes of performance metrics used to evaluate ML imputation methods: post-imputation bias of regression estimates, post-imputation predictive accuracy, and substantive model-free metrics. As an illustration, we applied data from a real-world breast cancer survival study. This comprehensive approach aimed to provide a thorough assessment of the effectiveness and reliability of ML imputation methods across various analytical perspectives. A simulated dataset with 30% Missing At Random (MAR) values was used. A number of single imputation (SI) methods - specifically KNN, missMDA, CART, missForest, missRanger, missCforest - and multiple imputation (MI) methods - specifically miceCART and miceRF - were evaluated. The performance metrics used were Gower's distance, estimation bias, empirical standard error, coverage rate, length of confidence interval, predictive accuracy, proportion of falsely classified (PFC), normalized root mean squared error (NRMSE), AUC, and C-index scores. The analysis revealed that in terms of Gower's distance, CART and missForest were the most accurate, while missMDA and CART excelled for binary covariates; missForest and miceCART were superior for continuous covariates. When assessing bias and accuracy in regression estimates, miceCART and miceRF exhibited the least bias. Overall, the various imputation methods demonstrated greater efficiency than complete-case analysis (CCA), with MICE methods providing optimal confidence interval coverage. In terms of predictive accuracy for Cox models, missMDA and missForest had superior AUC and C-index scores. Despite offering better predictive accuracy, the study found that SI methods introduced more bias into the regression coefficients compared to MI methods. This study underlines the importance of selecting appropriate imputation methods based on study goals and data types in time-to-event research. The varying effectiveness of methods across the different performance metrics studied highlights the value of using advanced machine learning algorithms within a multiple imputation framework to enhance research integrity and the robustness of findings. [ABSTRACT FROM AUTHOR]

Probability-arisen models play a considerable role in preparing a crucial stage for decision-making concerning reliability, engineering, and more closely related scenarios. Bearing in mind the consequential roles of probability-arisen models, we introduce and implement a new probabilistic model that has arisen by using the sine function, namely, the sine very flexible Weibull (SVF-Weibull) distribution. The proposed SVF-Weibull distribution is a result of a combination of the very flexible Weibull distribution with the sine-based strategy. For the SVF-Weibull distribution, point estimates are obtained. The assessment of the point estimates of the SVF-Weibull distribution is done via a simulation study. Finally, the consequential role of the SVF-Weibull distribution, illustrated by considering reliability and music engineering data sets. Furthermore, we implement some machine learning tools for predicting the reliability and music engineering data sets. The performances of the machine learning tools are assessed across many hidden variables. Our findings suggest that the artificial neural network method is more optimal than other methods for predicting the reliability and music engineering data sets. [ABSTRACT FROM AUTHOR]

This study compares the calculation of snow water equivalent (SWE) using machine learning algorithms with the conventional degree-day method. The study uses machine learning techniques such as LASSO, Random Forest, Support Vector Machines, and CatBoost. It proposes an innovative use of feature engineering (FE) to improve the accuracy and robustness of SWE predictions by machine learning intended for interpolation, extrapolation, or imputation of missing data. The performance of machine learning approaches is evaluated against the traditional degree-day method for predicting SWE. The study emphasizes and demonstrates gains when modeling is enhanced by transforming basic, raw data through feature engineering. The results, verified in a case study from the mountainous region of Slovakia, suggest that machine learning, particularly CatBoost with feature engineering, shows better results in SWE estimation in comparison with the degree-day method, although the authors present a refined application of the degree-day method by utilizing genetic algorithms. Nevertheless, the study finds that the degree-day method achieved accuracy with a Nash–Sutcliffe coefficient of efficiency NSE = 0.59, while the CatBoost technique enhanced with the proposed FE achieved an accuracy NSE = 0.86. The results of this research contribute to refining snow hydrology modeling and optimizing SWE prediction for improved decision-making in snow-dominated regions. [ABSTRACT FROM AUTHOR]

Due to the excessive growth of PM 2.5 in aerosol, the cases of lung cancer are increasing rapidly and are most severe among other types as the highest mortality rate. In most of the cases, lung cancer is detected with least symptoms at its later stage. Hence, clinical records may play a vital role to diagnose this disease at the correct stage for suitable medication to cure it. To detect lung cancer an accurate prediction method is needed which is significantly reliable. In the digital clinical record era with advancement in computing algorithms including machine learning techniques opens an opportunity to ease the process. Various machine learning algorithms may be applied over realistic clinical data but the predictive power is yet to be comprehended for accurate results. This paper envisages to compare twelve potential machine learning algorithms over clinical data with eleven symptoms of lung cancer along with two major habits of patients to predict a positive case accurately. The result has been found based on classification and heat map correlation. K-Nearest Neighbor Model and Bernoulli Naive Bayes Model are found most significant methods for early lung cancer prediction. [ABSTRACT FROM AUTHOR]

This study explores the performance of gradient boosting methods in bankruptcy prediction for a highly imbalanced dataset. We developed different heterogenous ensemble models based on three popular gradient boosting methods—XGBoost, LightGBM, and CatBoost. Our ensemble models were optimized using the cross‐validation method and the results of the hold‐out test sets showed that the optimized ensemble models not only outperform their base learners, but also improve the state‐of‐the‐art benchmark results on the same dataset. Interestingly, we observed that the data oversampling technique that is commonly used to address the class imbalance issue had an adverse impact on our ensemble models' performance. This indicates that our models are robust to the imbalanced dataset problem that typically degrades the classification performance of machine learning models. [ABSTRACT FROM AUTHOR]

In the field of data engineering in machine learning (ML), a crucial component is the process of scaling, normalization, and standardization. This process involves transforming data to make it more compatible with modeling techniques. In particular, this transformation is essential to ensure the suitability of the data for subsequent analysis. Despite the application of many conventional and relatively new approaches to ML, there remains a conspicuous lack of research, particularly in the geotechnical discipline. In this study, ML-based prediction models (i.e., RF, SVR, Cubist, and SGB) were developed to estimate the undrained shear strength (UDSS) of cohesive soil from the perspective of a wide range of data-scaling and transformation methods. Therefore, this work presents a novel ML framework based on data engineering approaches and the Cubist regression method to predict the UDSS of cohesive soil. A dataset including six different features and one target variable were used for building prediction models. The performance of ML models was examined considering the impact of the data pre-processing issue. For that purpose, data scaling and transformation methods, namely Range, Z-Score, Log Transformation, Box-Cox, and Yeo-Johnson, were used to generate the models. The results were then systematically compared using different sampling ratios to understand how model performance varies as various data scaling/transformation methods and ML algorithms were combined. It was observed that data transformation or data sampling methods had considerable or limited effects on the UDSS model performance depending on the algorithm type and the sampling ratio. Compared to RF, SVR, and SGB models, Cubist models provided higher performance metrics after applying the data pre-processing steps. The Box-Cox transformed Cubist model yielded the best prediction performance among the other models with an R2 of 0.87 for the 90% training set. Also, the UDSS prediction model generally yielded the best performance metrics when it was used with the transformed-based models (i.e., Box-Cox, Log, and Yeo-Johnson) than that of scaled-based (i.e., Range and Z-Score) models. The results show that the Cubist model has a higher potential for UDSS prediction, and data pre-processing methods have impacts on the predictive capacity of the evaluated regression models. [ABSTRACT FROM AUTHOR]

Crop models are instrumental in simulating resource utilization in agriculture, yet their complexity necessitates extensive calibration, which can impact the accuracy of yield predictions. Machine learning shows promise for enhancing yield estimations but relies on vast amounts of training data. This study aims to improve the pakchoi yield prediction accuracy of simulation models. We developed a stacking ensemble learning model that integrates three base models—EU-Rotate_N, Random Forest Regression and Support Vector Regression—with a Multi-layer Perceptron as the meta-model for the pakchoi dry matter yield prediction. To enhance the training dataset and bolster machine learning performance, we employed the EU-Rotate_N model to simulate daily dry matter yields for unsampled data. The test results revealed that the stacking model outperformed each base model. The stacking model achieved an R² value of 0.834, which was approximately 0.1 higher than that of the EU-Rotate_N model. The RMSE and MAE were 0.283 t/ha and 0.196 t/ha, respectively, both approximately 0.6 t/ha lower than those of the EU-Rotate_N model. The performance of the stacking model, developed with the expanded dataset, showed a significant improvement over the model based on the original dataset. [ABSTRACT FROM AUTHOR]

Beans are the main direct source of protein consumed by humans in the world and their productivity is directly linked to nitrogen. The short crop cycle imposes the need for fast methodologies for N quantification. In this work, we evaluated the performance of four machine learning algorithms in nitrogen estimation using NIR spectroscopy, comparing predictions between complete spectral data and only intervals obtained with the variable importance in projection (VIP). Doses of 0, 50, 100, and 150 kg ha−1 of N were applied and leaf reflectance was collected. Weka software was used to test the algorithms. The selection of the most effective spectral zones was made with the variable importance in projection (VIP). The intervals of 700–740 nm and 983–995 nm were considered the most important for the study of nitrogen. More efficient predictions were verified for RF and KNN models (R2 = 0.89, RMSE = 2.23 g kg−1; and R2 = 0.80, RMSE = 2.89 g kg−1, respectively) when only the most important spectral regions were included. The efficiency of nitrogen prediction based on NIR reflectance combined with machine learning was verified, which can serve as an important tool in precision agriculture. [ABSTRACT FROM AUTHOR]

In the whole world, approximately two billion people are using smartphones for their better life The smartphone's sensor data consist of high dimensions which increase complexity of training model and degrade overall performance of a classifier. We investigated and analyzed the performance of different machine learning algorithms with various dimensionality reduction techniques on publicly available smartphone-based recognition of human activities and postural transitions dataset. The highest accuracy was obtained at set 1 using support vector machine (SVM) linear kernel (97.40%) followed by set 3 using eXtreme boosting (XGBoost) (96.92%), set 2 using adaptive boosting multi-class (AdaBoost.M1) (95.69%), set 4 using AdaBoost. M1 (82.17%). Using different dimensionality reduction methods for set 1, principal component analysis (PCA) provided the highest accuracy using average neural network (AvNNET) (98.27%), while chi-squared gives the accuracy of 93.72% using the XGBoost classifier. For set 2, PCA gives the highest accuracy of 95.83% using the AvNNET classifier while for set 3 Chi-squared gives an accuracy of 84.46% using average neural network (AvNNET), SVM.L, and stochastic Gradient Boosting Machine (GBM) classifiers. The results reveal that overall dimensionality reduction methods outer performed. The outcomes will be helpful in making wise decisions in detecting various activities in anti-terrorists, surveillance, etc. [ABSTRACT FROM AUTHOR]

In recent years, healthcare monitoring systems (HMS) have increasingly integrated the Internet of Medical Things (IoMT) with cloud computing, leading to challenges related to data latency and efficient processing. This paper addresses these issues by introducing a Machine Learning-based Medical Data Segmentation (ML-MDS) approach that employs a k-fold random forest technique for efficient health data classification and latency reduction in a fog-cloud environment. Our method significantly improves latency issues, enhancing the Quality of Service (QoS) in healthcare systems and demonstrating its adaptability in heterogeneous network scenarios. We specifically employ the Random Forest algorithm to mitigate the common problem of overfitting in machine learning models, ensuring broader applicability across various healthcare contexts. Additionally, by optimizing data processing in fog computing layers, we achieve a substantial reduction in overall latency between healthcare sensors and cloud servers. This improvement is evidenced through a comparative performance analysis with existing models. The proposed framework not only ensures secure and scalable management of IoMT health data but also incorporates a stochastic approach to mathematically formulate performance indicators for the HMS queuing model. This model effectively predicts system response times and assesses the computing resources required under varying workload conditions. Our simulation results show a classification accuracy of 92%, a 56% reduction in latency compared to existing models, and an overall enhancement in e-healthcare service quality. [ABSTRACT FROM AUTHOR]

Completion of remedial mathematics has been identified as one of the keys to college success. However, completion rates in remedial mathematics have been low and are of much debate across America. This study leverages machine learning techniques in trying to predict and understand completion rates in remedial mathematics. The purpose of this study is to build machine learning models that can predict students that are least likely to complete remedial mathematics and identify which factors are most influential when computing those predictions. The study discovers random forest as the highest performing model. Furthermore, the study reveals that the remedial course a student begins with, credit completion rate, math placement score, and high school G.P.A as the most influential predictors of completion rates. The study also offers future research directions, especially in how to improve the performance of the machine learning models. [ABSTRACT FROM AUTHOR]

In manufacturing, it is a fatal problem to stop production due to a failure of facilities. Therefore, it is important to predict defects in facilities in advance and proactively maintain them. With the recent development of machine learning and deep learning technologies, research on predictive maintenance is also becoming active, and the performance of machine learning and deep learning technologies varies depending on the dataset. However, the existing dataset simply exists in normal and defective states, and the state between defects and normal was not simulated. Therefore, in this paper, a dataset was constructed by collecting additional attention data containing signals between normal and defective, and a simple classification experiment was conducted to verify the effectiveness of the data. Using the Prognostic dataset proposed in this paper, it is expected to be able to move toward a true predictive maintenance system that performs actions in advance, not after an accident. [ABSTRACT FROM AUTHOR]

There are multiple scenarios in machine learning where the data used presents a heavy bias towards one of the classes. Evaluating the performance of machine learning models in such imbalanced scenarios proves to be difficult and challenging, as one of the classes is poorly represented in the data, and this class is often more relevant to the end-user. An abundance of performance metrics have been devised and commonly used in order to solve these specific problems, however, there is often a lack of common agreement on which metric is best and which to use in specific imbalanced scenarios. In this study, we experimentally study the impact of choosing one metric over another in the evaluation of a classifier for binary classification, as well as the effect of data characteristics such as class imbalance and noise on those metrics. Based on our extensive empirical analysis, we provide a set of easy-to-follow guidelines for which performance metric is best to use depending on the context of the problem. Specifically, we highlight the importance of using multiple different metrics which are fundamentally different in imbalanced domains, we also display results on why the usage of Davis' interpolation of the area under the precision-recall curve and the Matthews Correlation Coefficient metrics should be preferred over other similar metrics, as well as why the usage of geometric mean and F 1 score should be avoided in scenarios likely to present noise on the labels. [ABSTRACT FROM AUTHOR]

Machine learning models' performance in simulating monthly rainfall–runoff in subtropical regions has not been sufficiently investigated. In this study, we evaluate the performance of six widely used machine learning models, including Long Short-Term Memory Networks (LSTMs), Support Vector Machines (SVMs), Gaussian Process Regression (GPR), LASSO Regression (LR), Extreme Gradient Boosting (XGB), and the Light Gradient Boosting Machine (LGBM), against a rainfall–runoff model (WAPABA model) in simulating monthly streamflow across three subtropical sub-basins of the Pearl River Basin (PRB). The results indicate that LSTM generally demonstrates superior capability in simulating monthly streamflow than the other five machine learning models. Using the streamflow of the previous month as an input variable improves the performance of all the machine learning models. When compared with the WAPABA model, LSTM demonstrates better performance in two of the three sub-basins. For simulations in wet seasons, LSTM shows slightly better performance than the WAPABA model. Overall, this study confirms the suitability of machine learning methods in rainfall–runoff modeling at the monthly scale in subtropical basins and proposes an effective strategy for improving their performance. [ABSTRACT FROM AUTHOR]

The sensitive attribute selection requires a well-trained machine-learning model to avoid unauthorized access to sensitive data. A new hybrid approach Boosted Neural Sensitive Attribute selection (BNSAD) method is used to extract sensitive attributes from the authorized data. This method is used in various access control methods mainly our proposed Hierarchical Attribute Based Access Control (HABAC) system. The HABAC gives data privacy by ensuring authorized individuals or groups only access the data. The main aim of the proposed system is to give a new hybrid machine-learning algorithm for data privacy. The BNSAD algorithm identifies the patterns from data and gives the boosted trained model as an input to the neural network to predict the sensitive attribute. The performance of seven machine learning algorithms was identified initially. The top three performance metrics of the algorithms taken for the new algorithm construction. The 94%, 92%, and 90% accuracy of the BNSAD algorithm for three datasets is higher than the existing three selected top three algorithms. The algorithm is implemented on BLP, RBAC, ABAC, and HABAC access control models. The final model accuracy of 98%, 98%, and 95% is achieved by BNSAD with HABAC for three medical datasets to identify the sensitive attributes for enhanced privacy of the data. The performance metrics for the proposed HABAC system are high with the existing models. [ABSTRACT FROM AUTHOR]

Machine learning interatomic potential (MLIP) has been widely adopted for atomistic simulations. While errors and discrepancies for MLIPs have been reported, a comprehensive examination of the MLIPs' performance over a broad spectrum of material properties has been lacking. This study introduces an analysis process comprising model sampling, benchmarking, error evaluations, and multi-dimensional statistical analyses on an ensemble of MLIPs for prediction errors over a diverse range of properties. By carrying out this analysis on 2300 MLIP models based on six different MLIP types, several properties that pose challenges for the MLIPs to achieve small errors are identified. The Pareto front analyses on two or more properties reveal the trade-offs in different properties of MLIPs, underscoring the difficulties of achieving low errors for a large number of properties simultaneously. Furthermore, we propose correlation graph analyses to characterize the error performances of MLIPs and to select the representative properties for predicting other property errors. This analysis process on a large dataset of MLIP models sheds light on the underlying complexities of MLIP performance, offering crucial guidance for the future development of MLIPs with improved predictive accuracy across an array of material properties. [ABSTRACT FROM AUTHOR]

The aim of this study was to test the morphometric features affecting 20-m sprint performance in children at the first level of primary education using machine learning (ML) algorithms. In this study, 130 male and 152 female volunteers aged between 6 and 11 years were included. After obtaining demographic information of the participants, skinfold thickness, diameter and circumference measurements, and 20-m sprint performance were determined. The study conducted three distinct experiments to determine the optimal ML technique for predicting outcomes. Initially, the entire feature space was utilized for training the ML models to establish a baseline performance. In the second experiment, only significant features identified through correlation analysis were used for training and testing the models, enhancing the focus on relevant predictors. Lastly, Principal Component Analysis (PCA) was employed to reduce the feature space, aiming to streamline model complexity while retaining data variance. These experiments collectively aimed to evaluate different feature selection and dimensionality reduction techniques, providing insights into the most effective strategies for optimizing predictive performance in the given context. The correlation-based selected features (Age, Height, waist circumference, hip circumference, leg length, thigh length, foot length) has produced a minimum Mean Squared Error (MSE) value of 0.012 for predicting the sprint performance in children. The effective utilization of correlation analysis in the selection of relevant features for our regression model suggests that the features selected exhibit robust linear associations with the target variable and can be relied upon as predictors. [ABSTRACT FROM AUTHOR]

Abstract\nHighlightsThe present study analyzes and optimizes machining characteristics, including feed rate (fz), depth of cut (ap), cutting speed (Vc), cutter-coated material (Mtc) and cutting-edge radius (rt), impacting on surface roughness (Ra), material removal rate (MRR) and tool wear (VB) of 42CrMo4 steel during dry end-milling. A total of 108 experimental runs were conducted, focusing on Ra, VB and MRR as response parameters. The nano TiAlN PVD-coated tool yielded better results for Ra and VB than did the TiCN/Al2O3 MT-CVD-coated tool. Then, Ra, VB and MRR optimization was carried out simultaneously using a Non-Dominated Sorting Genetic Algorithm III (NSGA-III) and Machine Learning (ML) models. Pareto solutions were found to offer a range of values for the three performance objectives: Ra (0.315–0.556 µm), VB (12.33–32.48 µm) and MRR (0.44–3.58 cm3/min). A quantitative performance score (Ps) ranking index was calculated to rank Pareto solutions for practical case studies. Validation experiments were subsequently performed to affirm that the optimal solution fell within a reasonable error range, with MAPE of 9.58% for Ra, 9.25% for VB and 13.39% for MRR. The validation results underscore the versatility of this approach, suggesting its applicability to a wide array of machining optimization challenges.108 experiments were conducted considering cutting speed (Vc), axial depth of cut (ap), feed per tooth (fz), cutting-edge radius (rt) and cutting-coated material (Mtc);The effect of cutting characteristics on surface roughness (Ra), flank wear (VB) and material removal rate (MRR) of 42CrMo4 was revealed;Machine learning models were developed to predict Ra, VB and MRR accurately;The multiobjective NSGA-III technique was employed to optimize Ra, VB and MRR simultaneously;These Pareto solutions underwent meticulous evaluation using a quantitative ranking index—the performance score (PS)—which thoughtfully considered user preferences and reliability;Experiments were performed to validate the optimum solution within an acceptance range of errors;108 experiments were conducted considering cutting speed (Vc), axial depth of cut (ap), feed per tooth (fz), cutting-edge radius (rt) and cutting-coated material (Mtc);The effect of cutting characteristics on surface roughness (Ra), flank wear (VB) and material removal rate (MRR) of 42CrMo4 was revealed;Machine learning models were developed to predict Ra, VB and MRR accurately;The multiobjective NSGA-III technique was employed to optimize Ra, VB and MRR simultaneously;These Pareto solutions underwent meticulous evaluation using a quantitative ranking index—the performance score (PS)—which thoughtfully considered user preferences and reliability;Experiments were performed to validate the optimum solution within an acceptance range of errors; [ABSTRACT FROM AUTHOR]

Nitrogen (N) is an essential nutrient for plant healthy growth and development, determining crop performance and yield outcomes. However, less than 50% of the N fertilizer applied in modern agricultural systems is effectively utilized by crops in the year of planting, causing resource waste and environmental harm. To improve N use efficiency (NUE), it is crucial to prioritize root performance as it primarily controls the plants' ability to absorb and utilize soil nutrients. The robustness of the root system is also closely related to crop lodging, which is an important factor that hinders the absorption, assimilation, translocation and distribution of N in major field crops. This review attempts to comprehensively summarize recent advances in plant root biology in the context of climate change, emphasizing its important role in soil nutrient dynamics and crop lodging risk, as well as its comprehensive impact on NUE of field grain crops. We discussed the importance of balanced and precision nutrient management to improve NUE and highlighted the innovative role of machine learning in modern agriculture, including root phenotyping, crop lodging detection and nutrient management decisions. Overall, combining machine learning with root biology, along with other practices such as crop rotation, cover cropping, use of biostimulants, and balanced nutrient supply, can reduce lodging risk and promote N uptake and utilization, thereby simultaneously improving NUE, crop productivity and agroecosystem sustainability. This helps ultimately secure our food supply while minimizing the negative impact of N on the environment facing unprecedent climate change. [ABSTRACT FROM AUTHOR]

Data missing is a ubiquitous problem in real-world systems that adversely affects the performance of machine learning algorithms. Although many useful imputation methods are available to address this issue, they often fail to consider the information provided by both features and labels. As a result, the performance of these methods might be constrained. Furthermore, feature selection as a data quality improvement technique has been widely used and has demonstrated its efficiency. To overcome the limitation of imputation methods, we propose a novel algorithm that combines data imputation and feature selection to tackle classification problems for mixed data. Based on the mean and standard deviation of quantitative features and the selecting probabilities of unique values of categorical features, our algorithm constructs different imputation models for quantitative and categorical features. Particle swarm optimization is used to optimize the parameters of the imputation models and select feature subsets simultaneously. Additionally, we introduce a legacy learning mechanism to enhance the optimization capability of our method. To evaluate the performance of the proposed method, seven algorithms and twelve datasets are used for comparison. The results show that our algorithm outperforms other algorithms in terms of accuracy and F 1 score and has reasonable time overhead. [ABSTRACT FROM AUTHOR]

In recent years, machine learning has become a frequently used method for statistical estimation. Random forest regression, decision tree regression, support vector regression and polynomial regression are commonly used supervised machine learning methods. The most commonly used loss function in gradient descent during the optimization phase of these methods is the quadratic loss function, which estimates model parameters by minimizing the cost. The selection of an appropriate loss function is crucial for method selection. There are several loss functions in the literature, such as absolute loss, logarithmic loss and squared error loss. In this study, we propose the use of an inverted normal loss function, which is a finite loss function, to gain a new perspective on minimizing cost and measuring performance in machine learning regression problems. We assert that this loss function provides more accurate estimations of cost minimisation as compared to the quadratic loss function, which is an infinite loss function. This article presents a new approach based on the inverted normal loss function for optimization in regression and performance metrics in machine learning. The procedure and its advantages are illustrated using a simulation study. [ABSTRACT FROM AUTHOR]

Chromophoric dissolved organic matter (CDOM) is a mixture of various types of organic matter and a useful parameter for monitoring complex inland surface waters. Remote sensing has been widely utilized to detect CDOM in various studies; however, in many cases, the dataset is relatively imbalanced in a single region. To address these concerns, data were acquired from hyperspectral images, field reflection spectra, and field monitoring data, and the imbalance problem was solved using a synthetic minority oversampling technique (SMOTE). Using the on-site reflectance ratio of the hyperspectral images, the input variables Rrs (452/497), Rrs (497/580), Rrs (497/618), and Rrs (684/618), which had the highest correlation with the CDOM absorption coefficient aCDOM (355), were extracted. Random forest and light gradient boosting machine algorithms were applied to create a CDOM prediction algorithm via machine learning, and to apply SMOTE, low-concentration and high-concentration datasets of CDOM were distinguished by 5 m−1. The training and testing datasets were distinguished at a 75%:25% ratio at low and high concentrations, and SMOTE was applied to generate synthetic data based on the training dataset, which is a sub-dataset of the original dataset. Datasets using SMOTE resulted in an overall improvement in the algorithmic accuracy of the training and test step. The random forest model was selected as the optimal model for CDOM prediction. In the best-case scenario of the random forest model, the SMOTE algorithm showed superior performance, with testing R2, absolute error (MAE), and root mean square error (RMSE) values of 0.838, 0.566, and 0.777 m−1, respectively, compared to the original algorithm's test values of 0.722, 0.493, and 0.802 m−1. This study is anticipated to resolve imbalance problems using SMOTE when predicting remote sensing-based CDOM. It is expected to produce and implement a machine learning model with improved reliable performance. [ABSTRACT FROM AUTHOR]

Background Objective Design Methods Results Conclusion Clinical Relevance Identifying health problems in audio‐recorded patient–nurse communication is important to improve outcomes in home healthcare patients who have complex conditions with increased risks of hospital utilization. Training machine learning classifiers for identifying problems requires resource‐intensive human annotation.To generate synthetic patient–nurse communication and to automatically annotate for common health problems encountered in home healthcare settings using GPT‐4. We also examined whether augmenting real‐world patient–nurse communication with synthetic data can improve the performance of machine learning to identify health problems.Secondary data analysis of patient–nurse verbal communication data in home healthcare settings.The data were collected from one of the largest home healthcare organizations in the United States. We used 23 audio recordings of patient–nurse communications from 15 patients. The audio recordings were transcribed verbatim and manually annotated for health problems (e.g., circulation, skin, pain) indicated in the Omaha System Classification scheme. Synthetic data of patient–nurse communication were generated using the in‐context learning prompting method, enhanced by chain‐of‐thought prompting to improve the automatic annotation performance. Machine learning classifiers were applied to three training datasets: real‐world communication, synthetic communication, and real‐world communication augmented by synthetic communication.Average F1 scores improved from 0.62 to 0.63 after training data were augmented with synthetic communication. The largest increase was observed using the XGBoost classifier where F1 scores improved from 0.61 to 0.64 (about 5% improvement). When trained solely on either real‐world communication or synthetic communication, the classifiers showed comparable F1 scores of 0.62–0.61, respectively.Integrating synthetic data improves machine learning classifiers' ability to identify health problems in home healthcare, with performance comparable to training on real‐world data alone, highlighting the potential of synthetic data in healthcare analytics.This study demonstrates the clinical relevance of leveraging synthetic patient–nurse communication data to enhance machine learning classifier performances to identify health problems in home healthcare settings, which will contribute to more accurate and efficient problem identification and detection of home healthcare patients with complex health conditions. [ABSTRACT FROM AUTHOR]

The efficiency of mechanical crushing is a key metric for evaluating machinery performance. However, traditional contact-based methods for measuring this efficiency are unable to provide real-time data monitoring and can potentially disrupt the production process. In this paper, we introduce a non-contact measurement technique for mechanical crushing efficiency based on deep learning algorithms. This technique utilizes close-range imaging equipment to capture images of crushed particles and employs deeply trained algorithmic programs rooted in symmetrical logical structures to extract statistical data on particle size. Additionally, we establish a relationship between particle size and crushing energy through experimental analysis, enabling the calculation of crushing efficiency data. Taking cement crushing equipment as an example, we apply this non-contact measurement technique to inspect cement particles of different sizes. Using deep learning algorithms, we automatically categorize and summarize the particle size ranges of cement particles. The results demonstrate that the crushing efficiencies of ore crushing particles, raw material crushing particles, and cement crushing particles can respectively reach 80.7%, 70.15%, and 80.27%, which exhibit a high degree of consistency with the rated value of the samples. The method proposed in this paper holds significant importance for energy efficiency monitoring in industries that require mechanical crushing. [ABSTRACT FROM AUTHOR]

With the increasing volume of data on the web, distributed file systems such as HDFS are widely used to provide efficient, fast and reliable access to the ever-growing volumes of data. Parameter tuning, a core performance management task, directly affects the performance of storage systems. However, large, non-linear, discrete and non-digital storage system configuration parameters significantly increase the storage system parameter tuning burden. Traditional manual tuning methods rely on a lot of a priori knowledge and have high tuning costs. In this paper, we propose a Smart gambler (SG) algorithm based on a simulated environment and Hill Climbing to reduce the probability of falling into a local optimum. In comparative experiments on publicly available datasets, SG achieves an average improvement of 3.14 % and 5.02 % in the highest tuning value over search algorithms such as Genetic Algorithm and an average time saving of 32.1 % and 50.6 % in tuning time for both workloads, respectively. [ABSTRACT FROM AUTHOR]

Despite huge effort over the years, the design of functionalised nanocarriers (NCs) for targeted drug delivery to endothelial cells is still to be completely unveiled. Dissipative particle dynamics (DPD) simulation is used to study the adhesion of NCs to endothelial cells under the influence of series of parameters such as shape, size, and ligand density of the NCs. However, preparing a performance space mapping that illustrates the penetration depths of NCs as a function of variations in their properties requires simulations of all possible NCs with the above-mentioned properties, which is not feasible. This challenge was addressed by leveraging a Gaussian process regression (GPR)-informed active learning strategy and an extensive exploration of numerous samples, each representing different properties of NCs. The performance space mapping reveals that NCs with rod and disc shapes exhibit superior penetration capabilities compared to those with a spherical shape. Furthermore, it demonstrates that smaller-sized rod-shaped NCs and larger-sized disc-shaped NCs tend to achieve better penetration. When considering smaller NCs, the influence of ligand density appears to be limited. On the contrary, for larger NCs, an increase in ligand density correlates with greater penetration depth, underscoring its substantial role in shaping their penetration capabilities. [ABSTRACT FROM AUTHOR]

Copyright of Przegląd Elektrotechniczny is the property of Przeglad Elektrotechniczny and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The paper addresses the issue of classification machine learning algorithm performance based on a novel probabilistic confusion matrix concept. The paper develops a theoretical framework which associates the proposed confusion matrix and the resulting performance metrics with the regular confusion matrix. The theoretical results are verified based on a wide variety of real-world classification problems and state-of-the-art machine learning algorithms. Based on the properties of the probabilistic confusion matrix, the paper then highlights the benefits of using the proposed concept both during the training phase and the application phase of a classification machine learning algorithm. [ABSTRACT FROM AUTHOR]

The rock classification system was initially applied to drill-and-blast tunnels and subsequently adapted for TBM tunnels; however, the majority of these systems primarily focused on rock stability while neglecting considerations of boreability. Compared with conventional tunnels, TBM tunnels are characterized by their rapid tunneling speed and excellent self-stabilization of the rock mass. Therefore, it is imperative to develop a novel rock mass classification system that considers both the tunneling efficiency of TBMs and the required support measures for tunnel construction. This paper introduces a novel rock classification system for TBM tunnels that accurately predicts the construction rate by evaluating the penetration rate and daily utilization, enabling a more precise assessment of daily advancement in tunneling. Firstly, the penetration rate and construction utilization in different rock strata are analyzed based on comprehensive statistics of existing construction data. Consequently, a discriminant matrix for classifying the boreability of rock is derived. Subsequently, employing the Ensemble Classifier method, a regression prediction model for rock boreability classification is established by incorporating input parameters such as thrust, torque, rotational speed, field penetration index, and the uniaxial compressive strength of rock. The validity of the proposed model is verified by comparing predicted machine performance with actual data sets. The proposed method presents a novel approach for predicting the performance of TBM construction. [ABSTRACT FROM AUTHOR]

This study investigates the use of developed machine learning techniques for modeling the performance of AlHayer, Saudi Arabia, wastewater treatment plant (ALWTP). Three physio-chemical characteristics were measured and predicted, including chemical oxygen demand (COD), biological oxygen demand (BOD), and suspended solids (SS), at ALWTP. The pre-evaluation of collected data revealed the effective capabilities of ALWTP for the removal of suspended solids, organic, and nutrient pollutants. To estimate the physio-chemical characteristics of ALWATP, four developed machine learning techniques were evaluated and compared. Logistic regression (LR), random forest (RF), gradient boosting (GB), and support vector regression (SVR) were designed. The evaluation of the proposed models showed RF outperformed other proposed models for estimating COD and SS with accuracy 91 % and 95 % in terms coefficient of determination (R2 ); however, GB was found the best, with accuracy 92 %, for detecting the BOD performance of ALWATP. This indicates ensemble learning models, RF and GB, can be considered a superiority soft solution for estimating physio-chemical characteristics of wastewater treatment plant. [ABSTRACT FROM AUTHOR]

Purpose: Predicting recurrence following stereotactic body radiotherapy (SBRT) for non‐small cell lung cancer provides important information for the feasibility of the individualized radiotherapy and allows to select the appropriate treatment strategy based on the risk of recurrence. In this study, we evaluated the performance of both machine learning models using positron emission tomography (PET) and computed tomography (CT) radiomic features for predicting recurrence after SBRT. Methods: Planning CT and PET images of 82 non‐small cell lung cancer patients who performed SBRT at our hospital were used. First, tumors were delineated on each CT and PET of each patient, and 111 unique radiomic features were extracted, respectively. Next, the 10 features were selected using three different feature selection algorithms, respectively. Recurrence prediction models based on the selected features and four different machine learning algorithms were developed, respectively. Finally, we compared the predictive performance of each model for each recurrence pattern using the mean area under the curve (AUC) calculated following the 0.632+ bootstrap method. Results: The highest performance for local recurrence, regional lymph node metastasis, and distant metastasis were observed in models using Support vector machine with PET features (mean AUC = 0.646), Naive Bayes with PET features (mean AUC = 0.611), and Support vector machine with CT features (mean AUC = 0.645), respectively. Conclusions: We comprehensively evaluated the performance of prediction model developed for recurrence following SBRT. The model in this study would provide information to predict the recurrence pattern and assist in making treatment strategies. [ABSTRACT FROM AUTHOR]

Attention is a cognitive process that is essential for human performance. Human development frequently depends on attention; however, this topic still requires additional research. Recently, EEG brain waves have been utilised to identify a person's attention states. This work aims to review numerous machine learning algorithms to analyse the Electroencephalographic (EEG) data to recognise human attention. Various machine learning approaches for the analysis of emotional states with EEG data were reviewed. Moreover, the analysis includes the performance achieved in various works; their benefits and disadvantages are reviewed. An EEG data processing pipeline and a review of human attention recognition are developed to evaluate the performance of various machine learning techniques. According to the analysis, the neural network framework achieves 99.81% accuracy. The results serve as a design framework for future systems using EEG data on brain activity to monitor people's health. [ABSTRACT FROM AUTHOR]

The high-dimensional nature of medical data frequently results in suboptimal performance of machine learning models. Applying feature selection before classification is necessary to improve the performance of classifiers. Although evolutionary-based wrapper feature selection methods are acknowledged for their superior performance in exploring optimal feature subsets, they have been demonstrated to carry the risk of overfitting and a potential loss of efficient search capability in the later stages of evolution. To address these issues, we propose a generalized wrapper feature selection method called Two Populations Based Feature Selection (TPBFS), which incorporates dual populations evolving in reverse directions to improve convergence speed. It introduces a probability-based crossover operation to mitigate overfitting and a record list to systematically track and replace optimal individuals, which helps to avoid getting stuck in local optima during later stages of evaluation. The experimental results demonstrate that TPBFS is effective in reducing the dimensionality of various medical datasets while guaranteeing the performance of classifiers. [ABSTRACT FROM AUTHOR]

Optimizing text preprocessing and text classification algorithms is an important, everyday task in large organizations and companies and it usually involves a labor-intensive and time-consuming effort. For example, the filtering and sorting of a large number of electronic mails (emails) are crucial to keeping track of the received information and converting it automatically into useful and profitable knowledge. Business emails are often unstructured, noisy, and with many abbreviations and acronyms, which makes their handling a challenging procedure. To overcome those challenges, a two-step classification approach is proposed, along with a two-cycle labeling procedure in order to speed up the labeling process. Every step incorporates a heuristic classification approach to assign emails to predefined classes by comparing several classification and text vectorization algorithms. These algorithms are compared and evaluated using the F1 score and balanced accuracy. The implementation of the proposed algorithm is demonstrated in a shipbroker agent operating in Greece with excellent performance, improving organization and administration while reducing expenses. [ABSTRACT FROM AUTHOR]

This review paper provides a comprehensive overview of machine vision pose measurement algorithms. The paper focuses on the state-of-the-art algorithms and their applications. The paper is structured as follows: the introduction in provides a brief overview of the field of machine vision pose measurement. Describes the commonly used algorithms for machine vision pose measurement. Discusses the factors that affect the accuracy and reliability of machine vision pose measurement algorithms. Summarizes the paper and provides future research directions. The paper highlights the need for more robust and accurate algorithms that can handle varying lighting conditions and occlusion. It also suggests that the integration of machine learning techniques may improve the performance of machine vision pose measurement algorithms. Overall, this review paper provides a comprehensive overview of machine vision pose measurement algorithms, their applications, and the factors that affect their accuracy and reliability. It provides a valuable resource for researchers and practitioners working in the field of computer vision. [ABSTRACT FROM AUTHOR]

In this study, we employed various machine learning models to predict metabolic phenotypes, focusing on thyroid function, using a dataset from the National Health and Nutrition Examination Survey (NHANES) from 2007 to 2012. Our analysis utilized laboratory parameters relevant to thyroid function or metabolic dysregulation in addition to demographic features, aiming to uncover potential associations between thyroid function and metabolic phenotypes by various machine learning methods. Multinomial Logistic Regression performed best to identify the relationship between thyroid function and metabolic phenotypes, achieving an area under receiver operating characteristic curve (AUROC) of 0.818, followed closely by Neural Network (AUROC: 0.814). Following the above, the performance of Random Forest, Boosted Trees, and K Nearest Neighbors was inferior to the first two methods (AUROC 0.811, 0.811, and 0.786, respectively). In Random Forest, homeostatic model assessment for insulin resistance, serum uric acid, serum albumin, gamma glutamyl transferase, and triiodothyronine/thyroxine ratio were positioned in the upper ranks of variable importance. These results highlight the potential of machine learning in understanding complex relationships in health data. However, it's important to note that model performance may vary depending on data characteristics and specific requirements. Furthermore, we emphasize the significance of accounting for sampling weights in complex survey data analysis and the potential benefits of incorporating additional variables to enhance model accuracy and insights. Future research can explore advanced methodologies combining machine learning, sample weights, and expanded variable sets to further advance survey data analysis. [ABSTRACT FROM AUTHOR]

Background Objectives Method Results Conclusion Crowdfunding is increasingly favoured by entrepreneurs for online financing. Predicting crowdfunding success can provide valuable guidance for stakeholders. It is a new attempt to evaluate the relative performance of different machine learning algorithms for crowdfunding prediction.This study aims to identify the key factors of crowdfunding, and find the different performance and usage of machine learning algorithms for crowdfunding prediction.We crawled data from MoDian.com, a Chinese crowdfunding platform, and predicted the crowdfunding performance using four machine learning algorithms, which is a new exploration in this area. Most of the existing literature focuses on empirical analysis. This work solves the problem of predicting crowdfunding performance using a dataset with a minimal number of highly contributive features, which has higher accuracy compared to the regression analysis.The experiment results show that feature‐selection‐based machine learning models are effective and beneficial in crowdfunding prediction.Feature selection can significantly improve the prediction performance of the machine learning models. KNN achieved the best prediction results with five features: number of backers, target amount, number of project likes, number of project comments, and sponsor fans. The prediction accuracy was improved by 16%, the precision was improved by 13.23%, the recall was improved by 22.66%, the F‐score was improved by 18.48%, and the AUC was improved by 14.9%. [ABSTRACT FROM AUTHOR]