Machine Learning-based Prediction Research Articles

Machine learning-based prediction of one-year mortality in ischemic stroke patients

Abstract Background Accurate prediction of mortality following an ischemic stroke is essential for tailoring personalized treatment strategies. This study evaluates the effectiveness of machine learning models in predicting one-year mortality after an ischemic stroke. Methods Five machine learning models were trained using data from a national stroke registry, with logistic regression demonstrating the highest performance. The SHapley Additive exPlanations (SHAP) analysis explained the model’s outcomes and defined the influential predictive factors. Results Analyzing 8183 ischemic stroke patients, logistic regression achieved 83% accuracy, 0.89 AUC, and an F1 score of 0.83. Significant predictors included stroke severity, pre-stroke functional status, age, hospital-acquired pneumonia, ischemic stroke subtype, tobacco use, and co-existing diabetes mellitus (DM). Discussion The model highlights the importance of predicting mortality in enhancing personalized stroke care. Apart from pneumonia, all predictors can serve the early prediction of mortality risk which supports the initiation of early preventive measures and in setting realistic expectations of disease outcomes for all stakeholders. The identified tobacco paradox warrants further investigation. Conclusion This study offers a promising tool for early prediction of stroke mortality and for advancing personalized stroke care. It emphasizes the need for prospective studies to validate these findings in diverse clinical settings.

Machine learning-based prediction models in medical decision-making in kidney disease: patient, caregiver, and clinician perspectives on trust and appropriate use.

This study aims to improve the ethical use of machine learning (ML)-based clinical prediction models (CPMs) in shared decision-making for patients with kidney failure on dialysis. We explore factors that inform acceptability, interpretability, and implementation of ML-based CPMs among multiple constituent groups. We collected and analyzed qualitative data from focus groups with varied end users, including: dialysis support providers (clinical providers and additional dialysis support providers such as dialysis clinic staff and social workers); patients; patients' caregivers (n = 52). Participants were broadly accepting of ML-based CPMs, but with concerns on data sources, factors included in the model, and accuracy. Use was desired in conjunction with providers' views and explanations. Differences among respondent types were minimal overall but most prevalent in discussions of CPM presentation and model use. Evidence of acceptability of ML-based CPM usage provides support for ethical use, but numerous specific considerations in acceptability, model construction, and model use for shared clinical decision-making must be considered. There are specific steps that could be taken by data scientists and health systems to engender use that is accepted by end users and facilitates trust, but there are also ongoing barriers or challenges in addressing desires for use. This study contributes to emerging literature on interpretability, mechanisms for sharing complexities, including uncertainty regarding the model results, and implications for decision-making. It examines numerous stakeholder groups including providers, patients, and caregivers to provide specific considerations that can influence health system use and provide a basis for future research.

Machine learning-based prediction of soil organic matter via smartphone

Color is an important property of soil that indicates soil composition and fertility. Soil organic matter (SOM) of darker color soils, which have rich humus and minerals, is higher than others such as red soils. The aim of this study is to construct models to predict SOM for a range of colors using a smartphone as an image-capturing device. Random forest of classification (RFC), random forest of logical regression (RFLR), convolutional neural network (CNN) and MobileNet models are compared, which is better for SOM prediction. Soil photos were collected by smartphone in a camera obscura with a steady light. After treatment by OpenCV, photos were separated by SOM content into five groups. The prediction accuracies of RFC, RFLR, CNN and MobileNet were 0.9743, 0.5614, 0.9600 and 0.7915, respectively. Based on these results, the RFC model for SOM has the best performance both in training and validation. The proposed combination of smartphone and machine learning-based prediction models provide a fast, economic, and robust approach to monitor, detect, and predict SOM contents in precision and intelligence agriculture.

Machine learning-based prediction of bioactivity in HIV-1 protease: insights from electron density analysis.

Aim: To develop a model for predicting the biological activity of compounds targeting the HIV-1 protease and to establish factors influencing enzyme inhibition.Materials & methods: Machine learning models were built based on a combination of Richard Bader's theory of Atoms in Molecules and topological analysis of electron density using experimental x-ray 'protein-ligand' complexes and inhibition constants data.Results & conclusion: Among all the models tested, logistic regression achieved the highest accuracy of 0.76 on the test set. The model's ability to differentiate between less active and highly active classes was relatively good, as indicated by an AUC-ROC score of 0.77. The analysis identified several critical factors affecting the biological activity of HIV-1 protease inhibitors, including the electron density contribution of hydrogen atoms, bond-critical points and particular amino acid residues. These findings provide new insights into how these molecular factors influence HIV-1 protease inhibition, emphasizing the importance of hydrogen bonding, glycine's flexibility and hydrophobic interactions in ligand binding.

A Forest Fire Prediction Model Based on Meteorological Factors and the Multi-Model Ensemble Method

More than half of South Korea’s land area is covered by forests, which significantly increases the potential for extensive damage in the event of a forest fire. The majority of forest fires in South Korea are caused by humans. Over the past decade, more than half of these types of fires occurred during the spring season. Although human activities are the primary cause of forest fires, the fact that they are concentrated in the spring underscores the strong association between forest fires and meteorological factors. When meteorological conditions favor the occurrence of forest fires, certain triggering factors can lead to their ignition more easily. The purpose of this study is to analyze the meteorological factors influencing forest fires and to develop a machine learning-based prediction model for forest fire occurrence, focusing on meteorological data. The study focuses on four regions within Gangwon province in South Korea, which have experienced substantial damage from forest fires. To construct the model, historical meteorological data were collected, surrogate variables were calculated, and a variable selection process was applied to identify relevant meteorological factors. Five machine learning models were then used to predict forest fire occurrence and ensemble techniques were employed to enhance the model’s performance. The performance of the developed forest fire prediction model was evaluated using evaluation metrics. The results indicate that the ensemble model outperformed the individual models, with a higher F1-score and a notable reduction in false positives compared to the individual models. This suggests that the model developed in this study, when combined with meteorological forecast data, can potentially predict forest fire occurrence and provide insights into the expected severity of fires. This information could support decision-making for forest fire management, aiding in the development of more effective fire response plans.

DeepCt: Predicting Pharmacokinetic Concentration-Time Curves and Compartmental Models from Chemical Structure Using Deep Learning.

After initial triaging using in vitro absorption, distribution, metabolism, and excretion (ADME) assays, pharmacokinetic (PK) studies are the first application of promising drug candidates in living mammals. Preclinical PK studies characterize the evolution of the compound's concentration over time, typically in rodents' blood or plasma. From this concentration-time (C-t) profiles, PK parameters such as total exposure or maximum concentration can be subsequently derived. An early estimation of compounds' PK offers the promise of reducing animal studies and cycle times by selecting and designing molecules with increased chances of success at the PK stage. Even though C-t curves are the major readout from a PK study, most machine learning-based prediction efforts have focused on the derived PK parameters instead of C-t profiles, likely due to the lack of approaches to model the underlying ADME mechanisms. Herein, a novel deep learning approach termed DeepCt is proposed for the prediction of C-t curves from the compound structure. Our methodology is based on the prediction of an underlying mechanistic compartmental PK model, which enables further simulations, and predictions of single- and multiple-dose C-t profiles.

Discovery of Alloy Catalysts for Ammonia Decomposition by Machine Learning-Based Prediction of Adsorption Energies

Ammonia decomposition has gained significant attention as an eco-friendly method for hydrogen production because it creates no carbon dioxide emissions. While Ru catalysts are known for their high activity in ammonia decomposition, their high cost makes them uneconomical for commercial use. Therefore, it is essential to explore novel alloy catalysts composed of inexpensive elements with high catalytic performance. Nitrogen adsorption energies serve as key descriptors indicating the catalytic performance for ammonia decomposition, and first-principle calculations can compute these energies. However, the screening of numerous alloy catalyst candidates through extensive first-principle calculations and experimental validations remains time-consuming due to the vast number of potential candidates. To address this, artificial intelligence and machine learning models are being developed to quickly predict catalyst performance, efficiently searching for promising catalyst candidates. In this study, we developed a machine-learning-based method to rapidly predict nitrogen adsorption energies using a graph-based artificial neural network, thereby efficiently searching for novel catalysts for ammonia decomposition. Our training dataset included the nitrogen adsorption energies of 30 pure transition metal catalyst candidates, as well as binary alloy catalyst candidates, including core-shell and intermetallic compounds. As a result, we successfully identified 12 catalyst candidates composed of inexpensive elements that are likely to exhibit catalytic performance comparable to Ru catalysts.

Machine learning-based prediction of scour depth evolution around spur dikes

ABSTRACT Spur dikes are pivotal elements in river training, serving to mitigate the dynamic alterations induced by river degradation and aggradation. Traditionally, scour prediction models have relied on regression techniques, but the advent of soft computing and machine learning has offered opportunities for enhanced accuracy. This study focuses on the development of hybrid machine-learning models, including eXtreme Gradient Boosting (XGBoost), random forest (RF), convolutional neural network–long short-term memory, and artificial neural network, optimized using genetic algorithms to predict both temporal scour depth variation and maximum scour depth around the initial spur dike in a series. The analysis reveals strong associations between scour depth and various parameters such as non-dimensional time, spacing, channel width, time-averaged velocity, and densimetric Froude number. The models are established through an iterative process involving four predictor combinations. Results demonstrate XGBoost as the top-performing model, consistently exhibiting superior performance with R2 of 0.99, root mean square error (RMSE) of 0.012, and mean absolute error of 0.008 during training, and R2 of 0.96, RMSE of 0.044, and Kling–Gupta efficiency of 0.98 during testing for predicting temporal scour depth. For non-dimensional maximum scour depth, it reached R2 of 0.99 and RMSE of 0.005 in training, with R2 > 0.91 across all combinations during testing. Although RF showcases commendable accuracy, it slightly lags in precision compared to XGBoost.

Machine Learning-Based Prediction of COVID-19 Patient Outcomes: Enhancing Clinical Decision-Making

Machine Learning-Based Prediction of COVID-19 Patient Outcomes: Enhancing Clinical Decision-Making

Machine Learning-Based Prediction for Incident Hypertension Based on Regular Health Checkup Data: Derivation and Validation in 2 Independent Nationwide Cohorts in South Korea and Japan.

Worldwide, cardiovascular diseases are the primary cause of death, with hypertension as a key contributor. In 2019, cardiovascular diseases led to 17.9 million deaths, predicted to reach 23 million by 2030. This study presents a new method to predict hypertension using demographic data, using 6 machine learning models for enhanced reliability and applicability. The goal is to harness artificial intelligence for early and accurate hypertension diagnosis across diverse populations. Data from 2 national cohort studies, National Health Insurance Service-National Sample Cohort (South Korea, n=244,814), conducted between 2002 and 2013 were used to train and test machine learning models designed to anticipate incident hypertension within 5 years of a health checkup involving those aged ≥20 years, and Japanese Medical Data Center cohort (Japan, n=1,296,649) were used for extra validation. An ensemble from 6 diverse machine learning models was used to identify the 5 most salient features contributing to hypertension by presenting a feature importance analysis to confirm the contribution of each future. The Adaptive Boosting and logistic regression ensemble showed superior balanced accuracy (0.812, sensitivity 0.806, specificity 0.818, and area under the receiver operating characteristic curve 0.901). The 5 key hypertension indicators were age, diastolic blood pressure, BMI, systolic blood pressure, and fasting blood glucose. The Japanese Medical Data Center cohort dataset (extra validation set) corroborated these findings (balanced accuracy 0.741 and area under the receiver operating characteristic curve 0.824). The ensemble model was integrated into a public web portal for predicting hypertension onset based on health checkup data. Comparative evaluation of our machine learning models against classical statistical models across 2 distinct studies emphasized the former's enhanced stability, generalizability, and reproducibility in predicting hypertension onset.

Accurate prediction of protein–ligand interactions by combining physical energy functions and graph-neural networks

We introduce an advanced model for predicting protein–ligand interactions. Our approach combines the strengths of graph neural networks with physics-based scoring methods. Existing structure-based machine-learning models for protein–ligand binding prediction often fall short in practical virtual screening scenarios, hindered by the intricacies of binding poses, the chemical diversity of drug-like molecules, and the scarcity of crystallographic data for protein–ligand complexes. To overcome the limitations of existing machine learning-based prediction models, we propose a novel approach that fuses three independent neural network models. One classification model is designed to perform binary prediction of a given protein–ligand complex pose. The other two regression models are trained to predict the binding affinity and root-mean-square deviation of a ligand conformation from an input complex structure. We trained the model to account for both deviations in experimental and predicted binding affinities and pose prediction uncertainties. By effectively integrating the outputs of the triplet neural networks with a physics-based scoring function, our model showed a significantly improved performance in hit identification. The benchmark results with three independent decoy sets demonstrate that our model outperformed existing models in forward screening. Our model achieved top 1% enrichment factors of 32.7 and 23.1 with the CASF2016 and DUD-E benchmark sets, respectively. The benchmark results using the LIT-PCBA set further confirmed its higher average enrichment factors, emphasizing the model’s efficiency and generalizability. The model’s efficiency was further validated by identifying 23 active compounds from 63 candidates in experimental screening for autotaxin inhibitors, demonstrating its practical applicability in hit discovery.Scientific contributionOur work introduces a novel training strategy for a protein–ligand binding affinity prediction model by integrating the outputs of three independent sub-models and utilizing expertly crafted decoy sets. The model showcases exceptional performance across multiple benchmarks. The high enrichment factors in the LIT-PCBA benchmark demonstrate its potential to accelerate hit discovery.

Machine Learning-Based Prediction of the Adsorption Characteristics of Biochar from Waste Wood by Chemical Activation.

This study employs machine learning models to predict the adsorption characteristics of biochar-activated carbon derived from waste wood. Activated carbon is a high-performance adsorbent utilized in various fields such as air purification, water treatment, energy production, and storage. However, its characteristics vary depending on the activation conditions or raw materials, making explaining or predicting them challenging using physicochemical or mathematical methods. Therefore, using machine learning techniques to determine the adsorption characteristics of activated carbon in advance will provide economic and time benefits for activated carbon production. Datasets, consisting of 108 points, were used to predict the adsorption characteristics of biochar-activated carbon derived from waste wood. The input variables were the activation conditions, and the iodine number of activated carbon was used as the output variable. The datasets were randomly split into 75% for training and 25% for model validation and normalized by the min-max function. Four models, including artificial neural networks, random forests, extreme gradient boosting, and support vector machines, were used to predict the adsorption properties of biochar-activated carbon. After optimization, the artificial neural network model was identified as the best model, with the highest coefficient determination (0.96) and the lowest mean squared error (0.004017). As a result of the SHAP analysis, activation time was the most crucial variable influencing the adsorption properties. The machine learning model precisely predicts the adsorption characteristics of biochar-activated carbon and can optimize the activated carbon production process.

Machine learning-based prediction of seismic response of elevated steel tanks

Seismic response analysis assesses structural safety and integrity during earthquakes and guides design for optimal performance, compliance with regulations, and risk mitigation. This study proposes a novel approach utilizing machine learning (ML) techniques to predict the seismic response of elevated steel tanks accurately. ML is a powerful tool that can handle complexity, improve accuracy, provide data-driven insights, and optimize designs for predicting seismic responses of structures. Moreover, this study evaluates various ML models containing Random Forest, Support Vector Regression, Adaptive Boosting, LightGBM, Histogram-based Gradient Boosting, XGBoost, CatBoost, Bagging Regression, and Decision Trees to identify the most effective predictive model. In general, the developed ML-based predictive models provide a promising tool for engineers involved in the seismic design and retrofitting of elevated steel tanks, aiding in optimizing designs and enhancing structural safety and resilience against seismic hazards. As a result, the HGB model obtained the most suitable performance compared to other developed models with higher R2 and lower RMSE. In addition, a sensitivity analysis was performed based on SHAP values, and it was found that the height of the liquid and the intensity measure were the most influential variables in the seismic response of the tank.

Study on Machine Learning-based Prediction of Lime Dosage for Mine Drainage Treatment

Study on Machine Learning-based Prediction of Lime Dosage for Mine Drainage Treatment

Machine learning-based prediction of single clad characteristics and non-destructive characterization of multi-layer deposited FeCoNiCrMo HEA on EN24 via laser cladding

The prediction of the proper metallurgical bonding and bead morphology in a laser cladding is very crucial to get a best profile of multi-clad layer. This work explores the laser cladding performance of FeCoCrNiMo high entropy alloy (HEA) on EN24 by analyzing the bead morphology with various process parameters. Machine learning (ML) technique is applied for two set of data where feasibility of the metallurgical bond is predicted and tested. From second set of data mainly three ML models, Support vector machine (SVM) radial kernels, K-Nearest neighbor (KNN), and multilayer perceptron (MLP) were applied to predict the outcomes. Sequential Minimal Optimization (SMO) algorithm is applied to identify the proper bonding between the substrate and cladded materials, which is classification-based ML model. The KNN model consistently attained the highest predictive accuracy across all response variables. It achieved perfect R² values and exhibited the lowest error metrics for clad width (w), clad height (h), clad depth (d), heat affected zone (HAZ), and Microhardness, establishing it as the most reliable and accurate model among the three applied ML model. To minimize the HAZ, the best conditions are lower laser powers, higher scanning speeds, and lower powder feed rate which is observed by contour plot. This reduces thermal diffusion into the substrate, maintaining material properties. Also, a non-destructive approach is used to characterize the multi-layer cladded before and after the deposition. The cladding process has significantly altered the microstructural and magnetic properties of the EN24 substrate, resulting in more refined grain structure and increased microhardness in the cladded region.

Experimental Study and Machine Learning-Based Prediction of the Abrasion Resistance of Manufactured Sand Concrete

To systematically analyze the impact of manufactured sand on the abrasion resistance of concrete, this paper investigates the correlation between sand type, sand ratio, stone powder content, compressive strength, and the abrasion resistance of manufactured sand concrete. Grey correlation analysis was conducted to assess the impact priority of each factor affecting the abrasion resistance, and prediction models for the abrasion resistance were developed using XGBoost, random forest, AdaBoost, and gradient boosting. The results indicate that compared to river sand concrete, C30 and C40 concrete prepared with limestone and diabase manufactured sand has 20% higher abrasion resistance due to the presence of stone powder and higher roughness and solidity. Within the range of 0.40 to 0.44, a lower sand ratio leads to higher abrasion resistance. For concrete prepared with manufactured sand containing 5% to 11% stone powder, the best abrasion resistance can be attained at a stone powder content of 9%, and microscopic analysis suggests the highest concrete density at this level. According to grey system theory, the influence of each affecting factor on the abrasion resistance follows the order: sand ratio > crushing value > roughness > compressive strength > stone powder content > 0.6. Compared to gradient boosting, random forest, and AdaBoost models, the XGBoost model adopted in this study showed relatively higher R2 and lower RMSE in both the training and testing sets, which proved its higher accuracy in predicting the abrasion loss of manufactured sand concrete. The machine learning models offer some guidance for predicting and enhancing the abrasion resistance of manufactured sand concrete in practical engineering.

Machine learning-based prediction of pulmonary embolism to reduce unnecessary computed tomography scans in gastrointestinal cancer patients: a retrospective multicenter study

This study aimed to develop a machine learning (ML) model for predicting pulmonary embolism (PE) in patients with gastrointestinal cancers, a group at increased risk for PE. We conducted a retrospective, multicenter study analyzing patients who underwent computed tomographic pulmonary angiography (CTPA) between 2010 and 2020. The study utilized demographic and clinical data, including the Wells score and D-dimer levels, to train a random forest ML model. The model’s effectiveness was assessed using the area under the receiver operating curve (AUROC). In total, 446 patients from hospital A and 139 from hospital B were included. The training set consisted of 356 patients from hospital A, with internal validation on 90 and external validation on 139 patients from hospital B. The model achieved an AUROC of 0.736 in hospital A and 0.669 in hospital B. The ML model significantly reduced the number of patients recommended for CTPA compared to the conventional diagnostic strategy (hospital A; 100.0% vs. 91.1%, P < 0.001, hospital B; 100.0% vs. 93.5%, P = 0.003). The results indicate that an ML-based prediction model can reduce unnecessary CTPA procedures in gastrointestinal cancer patients, highlighting its potential to enhance diagnostic efficiency and reduce patient burden.

Machine learning-based prediction of sarcopenia in community-dwelling middle-aged and older adults: findings from the CHARLS.

Sarcopenia is a prominent issue among aging populations and associated with poor health outcomes. This study aimed to examine the predictive value of questionnaire and biomarker data for sarcopenia, and to further develop a user-friendly calculator for community-dwelling middle-aged and older adults. We used two waves (2011 and 2013) of the China Health and Retirement Longitudinal Study (CHARLS) to predict sarcopenia, defined by the Asian Working Group for Sarcopenia 2019 criteria. We restricted the analytical sample to adults aged 45 or above (N = 2934). Five machine learning models were used to construct Q-based (only questionnaire variables), Bio-based (only biomarker variables), and combined (questionnaire plus biomarker variables) models. Area under the receiver operating characteristic curve (AUROC) was used for performance assessment. Temporal external validation was performed based on two datasets from CHARLS. Important predictors were identified by Shapley values and coefficients. Extreme gradient boosting (XGBoost), considering both questionnaire and biomarker characteristics, emerged as the optimal model, and its AUROC was 0.759 (95% CI: 0.747-0.771) at a decision threshold of 0.20 on the test set. Models also performed well on the external datasets. We found that cognitive function was the most important predictor in both Q-based and combined models, and blood urea nitrogen was the most important predictor in the Bio-based model. Other key predictors included education, haematocrit, total cholesterol, drinking, number of chronic diseases, and instrumental activities of daily living score. Our findings offer a potential for early screening and targeted prevention of sarcopenia among middle-aged and older adults in the community setting.

Meta-analysis of prediction models for predicting lymph node metastasis in thyroid cancer

BackgroundThe purpose of this systematic review and meta-analysis is to assess the efficacy of various machine learning (ML) techniques in predicting preoperative lymph node metastasis (LNM) in patients diagnosed with papillary thyroid carcinoma (PTC). Although prior studies have investigated the potential of ML in this context, the current evidence is not sufficiently strong. Hence, we undertook a thorough analysis to ascertain the predictive accuracy of different ML models and their practical relevance in predicting preoperative LNM in PTC patients.Materials and methodsIn our search, we thoroughly examined PubMed, Cochrane Library, Embase, and Web of Science, encompassing their complete database history until December 3rd, 2022. To evaluate the potential bias in the machine learning models documented in the included studies, we employed the Prediction Model Risk of Bias Assessment Tool (PROBAST).ResultsA total of 107 studies, involving 136,245 patients, were included. Among them, 21,231 patients showed central LNM (CLNM) and 4,637 had lateral LNM (LLNM). The meta-analysis results revealed that the c-index for predicting LNM, CLNM, and LLNM were 0.762 (95% CI: 0.747–0.777), 0.762 (95% CI: 0.747–0.777), and 0.803 (95% CI: 0.773–0.834) in the training set, and 0.773 (95% CI: 0.754–0.791), 0.762 (95% CI: 0.747–0.777), and 0.829 (95% CI: 0.779–0.879) in the validation set, respectively. A total of 134 machine learning-based prediction models were included, covering 10 different types. Logistic Regression (LR) was the most commonly used model, accounting for 81.34% (109/134) of the included models.ConclusionsMachine learning methods have shown a certain level of accuracy in predicting preoperative LNM in PTC patients, indicating their potential as a predictive tool. Their use in the clinical management of PTC holds great promise. Among the various ML models investigated, the performance of logistic regression-based nomograms was deemed satisfactory.

Medical diagnosis based on artificial intelligence and decision support system in the management of health development.

Medical diagnosis plays a critical role in our daily lives. Every day, over 10 billion cases of both mental and physical health disorders are diagnosed and reported worldwide. To diagnose these disorders, medical practitioners and health professionals employ various assessment tools. However, these tools often face scrutiny due to their complexity, prompting researchers to increase their experimental parameters to provide accurate justifications. Additionally, it is essential for professionals to properly justify, interpret, and analyse the results from these prediction tools. This research paper explores the use of artificial intelligence and advanced analytics in developing Clinical Decision Support Systems (CDSS). These systems are capable of diagnosing and detecting patterns of various medical disorders. Various machine learning algorithms contribute to building these assessment tools, with the Network Pattern Recognition (NEPAR) algorithm being the first to aid in developing CDSS. Over time, researchers have recognised the value of machine learning-based prediction models in successfully justifying medical diagnoses. The proposed CDSS models have demonstrated the ability to diagnose mental disorders with an accuracy of up to 89% using only 28 questions, without requiring human input. For physical health issues, additional parameters are used to enhance the accuracy of CDSS models. Consequently, medical professionals are increasingly relying on these machine learning-based CDSS models, utilising these tools to improve medical diagnosis and assist in decision-making. The different cross-validation values are considered to remove the data biasness.