Machine Model Research Articles

Acupuncture indication knowledge bases: meridian entity recognition and classification based on ACUBERT.

In acupuncture diagnosis and treatment, non-quantitative clinical descriptions have limited the development of standardized treatment methods. This study explores the effectiveness and the reasons for discrepancies in the entity recognition and classification of meridians in acupuncture indication using the Acupuncture Bidirectional Encoder Representations from Transformers (ACUBERT) model. During the research process, we selected 54 593 different entities from 82 acupuncture medical books as the pretraining corpus for medical literature, conducting classification research on Chinese medical literature using the BERT model. Additionally, we employed the support vector machine and Random Forest models as comparative benchmarks and optimized them through parameter tuning, ultimately leading to the development of the ACUBERT model. The results show that the ACUBERT model outperforms other baseline models in classification effectiveness, achieving the best performance at Epoch = 5. The model's "precision," "recall," and F1 scores reached above 0.8. Moreover, our study has a unique feature: it trains the meridian differentiation model based on the eight principles of differentiation and zang-fu differentiation as foundational labels. It establishes an acupuncture-indication knowledge base (ACU-IKD) and ACUBERT model with traditional Chinese medicine characteristics. In summary, the ACUBERT model significantly enhances the classification effectiveness of meridian attribution in the acupuncture indication database and also demonstrates the classification advantages of deep learning methods based on BERT in multi-category, large-scale training sets. Database URL: http://acuai.njucm.edu.cn:8081/#/user/login?tenantUrl=default.

Qualitative discrimination and quantitative prediction of salt in aqueous solution based on near-infrared spectroscopy

Freshwater resources have been gradually salinized in recent years, dramatically impacting the ecosystem and human health. Therefore, it is necessary to detect the salinity of freshwater resources. However, traditional detection methods make it difficult to check the type and concentration of salt quickly and accurately in solution. This paper uses a portable near-infrared spectrometer to qualitatively discriminate and quantitatively predict the salt in the solution. The study was carried out by adding ten salts of NaCl, KCl, MgCl2, CaCl2, Na2CO3, K2CO3, CaCO3, Na2SO4, K2SO4, MgSO4 to 2 milliliters of deionized water to prepare a single salt solution (0.02% - 1.00%) totaling 100 sets. It was found that the Support vector machine (SVM) model was only effective in discriminating the class of salt anions in the solution. The Partial least squares-discriminant analysis (PLS-DA) model, on the other hand, can effectively discriminate the classes of salt in solution, and the accuracies of the optimal model prediction set and the interactive validation set are 98.86% and 99.66%, respectively. Furthermore, the Partial least squares regression (PLSR) models can accurately predict the concentration of NaCl, KCl, MgCl2, CaCl2, Na2CO3, K2CO3, CaCO3, Na2SO4, K2SO4, MgSO4 salt solutions. The coefficients of determination R2 of their model interactive validation sets were 0.99, 0.99, 0.99, 0.97, 0.99, 0.99, 0.98, 0.99, 0.98, and 0.98, respectively. This study shows that NIRS can achieve rapid and accurate qualitative and quantitative detection of salts in solution, which provides technical support for the utilization of safe water resources.

Class imbalanced data handling with cyberattack classification using Hybrid Salp Swarm Algorithm with deep learning approach

Cyberattack classification involves applying deep learning (DL) and machine learning (ML) models to categorize digital threats based on their features and behaviors. These models examine system logs, network traffic, or other associated data patterns to discriminate between standard activities and malicious actions. Efficient cyberattack classification is vital for on-time threat detection and response, permitting cybersecurity specialists to categorize and reduce potential risks to a system. Handling class-imbalanced data in cyberattack classification using DL is critical for achieving exact and robust models. In cybersecurity databases, instances of normal behavior frequently significantly outnumber instances of cyberattacks, foremost due to biased methods that may complete poorly on minority classes. To address this issue approaches such as oversampling the lesser class, undersampling the popular class, or using more advanced systems can be used. These plans defend that the DL technique is more complex when determining cyberattacks, so it increases complete performance and adapts the effect of the imbalance class on the classification results. This study presents a novel Hybrid Salp Swarm Algorithm with a DL Approach for Cyberattack Classification (HSSADL-CAC) technique. The HSSADL-CAC method intends to resolve class imbalance data handling with an optimum DL model for the recognition of cyberattacks. At first, the HSSADL-CAC method experiences data normalization as a pre-processing stage. The HSSADL-CAC technique uses the ADASYN approach to handle class imbalance problems. In addition, the HSSADL-CAC technique applies an HSSA-based feature selection approach. The HSSADL-CAC technique detects cyberattacks using a deep extreme learning machine (DELM) model. Finally, the hyperparameter tuning of the ELM model takes place by utilizing the beluga whale optimization (BWO) model. The performance analysis of the HSSADL-CAC technique employs a benchmark database. The comprehensive comparison research indicates the superior performance of the HSSADL-CAC technique in the cyberattack detection procedure.

SnO2-Co3O4 nanocomposite sensor: Achieving ultra-selective hydrogen detection in mixed gas environments

This paper addresses the long-standing challenge of poor selectivity in MOS sensors towards gases with similar properties, exemplified by the difficulty in distinguishing between H2 and CO in mixed environments. In this paper, n-SnO2/p-Co3O4 nanocomposites have been prepared as sensing materials by a straightforward pyrolytic bimetallic organic framework method. Optimal hydrogen detection is achieved with a Sn: Co molar ratio of 1:0.15, demonstrating a high response (S=Ra/Rg=22.75 for 300 ppm H2) and exceptional anti-interference capability against CO (SH2/SCO=11.67) at 325°C. Furthermore, the incorporation of Extreme Learning Machine (ELM) model for predicting hydrogen concentration significantly reduces the interference effect common in mixed gases, achieving a high prediction accuracy a remarkably low relative error of just 0.6 %. The mechanism behind the enhanced sensitivity and selectivity of the sensor is elucidated in the paper, offering both theoretical insight and practical solution to the issue of selectivity in MOS sensors.

Development of a System for Predicting Hospitalization Time for Patients With Traumatic Brain Injury Based on Machine Learning Algorithms: User-Centered Design Case Study.

Currently, the treatment and care of patients with traumatic brain injury (TBI) are intractable health problems worldwide and greatly increase the medical burden in society. However, machine learning-based algorithms and the use of a large amount of data accumulated in the clinic in the past can predict the hospitalization time of patients with brain injury in advance, so as to design a reasonable arrangement of resources and effectively reduce the medical burden of society. Especially in China, where medical resources are so tight, this method has important application value. We aimed to develop a system based on a machine learning model for predicting the length of hospitalization of patients with TBI, which is available to patients, nurses, and physicians. We collected information on 1128 patients who received treatment at the Neurosurgery Center of the Second Affiliated Hospital of Anhui Medical University from May 2017 to May 2022, and we trained and tested the machine learning model using 5 cross-validations to avoid overfitting; 28 types of independent variables were used as input variables in the machine learning model, and the length of hospitalization was used as the output variables. Once the models were trained, we obtained the error and goodness of fit (R2) of each machine learning model from the 5 rounds of cross-validation and compared them to select the best predictive model to be encapsulated in the developed system. In addition, we externally tested the models using clinical data related to patients treated at the First Affiliated Hospital of Anhui Medical University from June 2021 to February 2022. Six machine learning models were built, including support vector regression machine, convolutional neural network, back propagation neural network, random forest, logistic regression, and multilayer perceptron. Among them, the support vector regression has the smallest error of 10.22% on the test set, the highest goodness of fit of 90.4%, and all performances are the best among the 6 models. In addition, we used external datasets to verify the experimental results of these 6 models in order to avoid experimental chance, and the support vector regression machine eventually performed the best in the external datasets. Therefore, we chose to encapsulate the support vector regression machine into our system for predicting the length of stay of patients with traumatic brain trauma. Finally, we made the developed system available to patients, nurses, and physicians, and the satisfaction questionnaire showed that patients, nurses, and physicians agreed that the system was effective in providing clinical decisions to help patients, nurses, and physicians. This study shows that the support vector regression machine model developed using machine learning methods can accurately predict the length of hospitalization of patients with TBI, and the developed prediction system has strong clinical use.

A missing data processing method for dam deformation monitoring data using spatiotemporal clustering and support vector machine model

Deformation monitoring is a critical measure for intuitively reflecting the operational behavior of a dam. However, the deformation monitoring data are often incomplete due to environmental changes, monitoring instrument faults, and human operational errors, thereby often hindering the accurate assessment of actual deformation patterns. This study proposed a method for quantifying deformation similarity between measurement points by recognizing the spatiotemporal characteristics of concrete dam deformation monitoring data. It introduces a spatiotemporal clustering analysis of the concrete dam deformation behavior and employs the support vector machine model to address the missing data in concrete dam deformation monitoring. The proposed method was validated in a concrete dam project, with the model error maintaining within 5%, demonstrating its effectiveness in processing missing deformation data. This approach enhances the capability of early-warning systems and contributes to enhanced dam safety management.

Prediction of Clinical Outcomes in Psychotic Disorders Using Artificial Intelligence Methods: A Scoping Review.

Psychotic disorders are major psychiatric disorders that can impact multiple domains including physical, social, and psychological functioning within individuals with these conditions. Being able to better predict the outcomes of psychotic disorders will allow clinicians to identify illness subgroups and optimize treatment strategies in a timely manner. In this scoping review, we aimed to examine the accuracy of the use of artificial intelligence (AI) methods in predicting the clinical outcomes of patients with psychotic disorders as well as determine the relevant predictors of these outcomes. This review was guided by the PRISMA Guidelines for Scoping Reviews. Seven electronic databases were searched for relevant published articles in English until 1 February 2024. Thirty articles were included in this review. These studies were mainly conducted in the West (63%) and Asia (37%) and published within the last 5 years (83.3%). The clinical outcomes included symptomatic improvements, illness course, and social functioning. The machine learning models utilized data from various sources including clinical, cognitive, and biological variables such as genetic, neuroimaging measures. In terms of main machine learning models used, the most common approaches were support vector machine, random forest, logistic regression, and linear regression models. No specific machine learning approach outperformed the other approaches consistently across the studies, and an overall range of predictive accuracy was observed with an AUC from 0.58 to 0.95. Specific predictors of clinical outcomes included demographic characteristics (gender, socioeconomic status, accommodation, education, and employment); social factors (activity level and interpersonal relationships); illness features (number of relapses, duration of relapses, hospitalization rates, cognitive impairments, and negative and disorganization symptoms); treatment (prescription of first-generation antipsychotics, high antipsychotic doses, clozapine, use of electroconvulsive therapy, and presence of metabolic syndrome); and structural and functional neuroimaging abnormalities, especially involving the temporal and frontal brain regions. The current review highlights the potential and need to further refine AI and machine learning models in parsing out the complex interplay of specific variables that contribute to the clinical outcome prediction of psychotic disorders.

Quantifying seasonal variations in pollution sources with machine learning-enhanced positive matrix factorization

As the pace of industrialization and urbanization accelerates, water quality management faces increasing challenges, with traditional methods for pollutant source apportionment often proving inadequate in handling complex environmental data. This study enhances the precision and reliability of pollutant source identification by integrating Positive Matrix Factorization (PMF) models with diverse machine learning techniques. Utilizing data from 17 water quality monitoring stations along the Wuding River from 2017 to 2021, we employed Random Forest (RF), Support Vector Machine (SVM), Elastic Net (EN), and Extreme Gradient Boosting (XGBoost) models to predict the Water Quality Index (WQI) during dry and wet seasons. Results indicate that the RF model exhibited optimal performance in the dry season (R2 = 0.93), while the SVM was superior in the wet season (R2 = 0.94). SHAP (SHapley Additive exPlanations) value analysis identified CODMn and NH3-N as significant influencers on WQI in the dry season, whereas COD, BOD, and TP gained prominence during the wet season. SHAP values reveal the contribution of each feature to the model output, thereby enhancing the model’s transparency and interpretability. Additionally, feature importance identified by machine learning was utilized as weights to optimize the contribution rates predicted by the PMF model. The optimised model was able to identify the contribution of domestic and farm effluent discharges more accurately in the dry season, with a significant increase in the percentage of identification from 19.4 % to 45.4 %, and an increase in the percentage of contribution from agricultural non-point sources and domestic effluent in the rainy season. This research offers a novel perspective on the characteristics of river water pollution and holds significant implications for formulating data-driven environmental management strategies.

An Integrated Hybrid Convolutional-Support Vector Machine (CSVM) Model with the CWT Method for Hearing Disorder Detection using AEP Signals

Hearing disorder is the most widespread sensory disability worldwide, impairing human communication and learning. An early and accurate hearing disorder detection system using an electroencephalogram (EEG) is the appropriate technique for dealing with this concern. The most significant modality for diagnosing hearing deficiency among EEG control signals is the auditory evoked potential (AEP), which is generated in the cortical region of the brain through auditory stimulus. This study aims to develop an efficient approach for detecting hearing disorders. For this purpose, this study has designed a hybrid model based on the convolutional operation of CNN and the SVM classifier. Initially, the CWT method was utilized to transform the raw AEP signals into time-frequency images. Then, the extracted features were classified using the proposed CSVM model. To test the robustness of the proposed model, this study also implemented a convolutional neural network (CNN) and support vector machine (SVM) with the same parameters. The experimental results with the hybrid CSVM model showed superior performance on the publicly available AEP dataset by achieving 94.48% testing accuracy, 96.40% precision, 92.96% recall, 94.65% F1 score, 88.95% Cohen Kappa score, which indicates that the proposed hybrid model could be used for early hearing disorder detection. Future enhancements will concentrate on identifying different hearing-signal-based data and the cloud-based, automated classification of AEP signals.

SSTtrack: A unified hyperspectral video tracking framework via modeling spectral-spatial-temporal conditions

Hyperspectral video contains rich spectral, spatial, and temporal conditions that are crucial for capturing complex object variations and overcoming the inherent limitations (e.g., multi-device imaging, modality alignment, and finite spectral bands) of regular RGB and multi-modal video tracking. However, existing hyperspectral tracking methods frequently encounter issues including data anxiety, band gap, huge volume, and weakness of the temporal condition embedded in video sequences, which result in unsatisfactory tracking capabilities. To tackle the dilemmas, we present a unified hyperspectral video tracking framework via modeling spectral-spatial-temporal conditions end-to-end, dubbed SSTtrack. First, we design the multi-modal generation adapter (MGA) to explore the interpretability benefits of combining physical and machine models for learning the multi-modal generation and bridging the band gap. To dynamically transfer and interact with multiple modalities, we then construct a novel spectral-spatial adapter (SSA). Finally, we design a temporal condition adapter (TCA) for injecting the temporal condition to guide spectral and spatial feature representations to capture static and instantaneous object properties. SSTtrack follows the prompt learning paradigm with the addition of few trainable parameters (0.575 M), resulting in superior performance in extensive comparisons. The code will be released at https://github.com/YZCU/SSTtrack.

Construction and validation of a bioinformatics‑based screen for cuproptosis‑related genes and risk model for Alzheimer's disease.

The present study aimed to validate the association between core cuproptosis genes (CRGs) and Alzheimer's disease (AD) from both bioinformatics and experimental perspectives and also to develop a risk prediction model. To this end, 78 human‑derived temporal back samples were analyzed from GSE109887, and the biological functions of the resulting CRGs were explored by cluster analysis, weighted gene co‑expression network analysis and similar methods to identify the best machine model. Moreover, an external dataset GSE33000 and a nomogram were used to validate the model. The mRNA and protein expression of CRGs were validated using the SH‑SY5Y cell model and the Sprague‑Dawley rat animal model. The RT‑qPCR and western blotting results showed that the mRNA and protein expression content of dihydrolipoamide dehydrogenase, ferredoxin 1, glutaminase and pyruvate dehydrogenase E1 subunit β decreased, and the expression of dihydrolipoamide branched chain transacylase E2 increased in AD, which supported the bioinformatic analysis results. The CRG expression alterations affected the aggregation and infiltration of certain immune cells. The present study also confirmed the accuracy and validity of AD diagnostic models and nomograms, and validated the association between five CRGs and AD, indicating a significant difference between patients with AD and healthy individuals. Therefore, CRGs are expected to serve as relevant biomarkers for the diagnosis and prognostic monitoring of AD.

Machine learning supported ultrasonic testing for characterization of cracks in polyethylene pipes

This study aims to develop a machine learning supported ultrasonic guided wave testing (UGWT) for detection and characterization of cracks in polyethylene (PE) pipes used for natural gas distribution. Ultrasonic testing is conducted to investigate the wave-crack interaction and determine damping properties of PE pipes. Finite element models are further built with material properties fine-tuned by wave attenuation experiments combined with cross-correlation analysis between the simulated and experimental signals. A synthetic database is populated using sensed signals over a wide range of crack geometries from numerical simulations. Finally, support vector machine (SVM) models are developed with different feature selections for classification of crack geometry and the accuracy is evaluated using both simulated and experimental cases. The findings demonstrate the potential of locating the circumferential position of crack with the ring-focusing method and classifying crack geometry in PE pipes using SVM model with central frequency features.

Diagnosis of acute hyperglycemia based on data-driven prediction models

Acute hyperglycemia is a common endocrine and metabolic disorder that seriously threatens the health and life of patients. Exploring effective diagnostic methods and treatment strategies for acute hyperglycemia to improve treatment quality and patient satisfaction is currently one of the hotspots and difficulties in medical research. This article introduced a method for diagnosing acute hyperglycemia based on data-driven prediction models. In the experiment, clinical data from 1000 patients with acute hyperglycemia were collected. Through data cleaning and feature engineering, 10 features related to acute hyperglycemia were selected, including BMI (Body Mass Index), TG (triacylglycerol), HDL-C (High-density lipoprotein cholesterol), etc. The support vector machine (SVM) model was used for training and testing. The experimental results showed that the SVM model can effectively predict the occurrence of acute hyperglycemia, with an average accuracy of 96 %, a recall rate of 84 %, and an F1 value of 89 %. The diagnostic method for acute hyperglycemia based on data-driven prediction models has a certain reference value, which can be used as a clinical auxiliary diagnostic tool to improve the early diagnosis and treatment success rate of acute hyperglycemia patients.

Global trends and hotspots in the study of the effects of PM2.5 on ischemic stroke

AimThe objective of this study was to visually analyse global research trends and hotspots regarding the role of PM2.5 in ischemic stroke.MethodsThe Web of Science core collection database was used to search the literature on PM2.5 and ischemic stroke from 2006 to 2024. Visualization analysis was conducted using CiteSpace, VOSviewer, and an online bibliometric platform.ResultsThe analysis comprises 190 articles published between 2006 and 2024 by 1229 authors from 435 institutions in 39 countries, across 78 journals. Wellenius GA has the highest number of published and cited papers. China has the highest number of papers, while Canada has the highest citation frequency. Capital Medical University published the highest number of papers, and Harvard University had the highest citation frequency for a single paper. The study investigated the impact of PM2.5 on ischemic stroke in three phases. The first phase analysed hospitalisation rates for correlations. The second phase utilised large-scale multi-cohort data from around the world. The third phase involved studying global exposure risk through machine learning and model construction. Currently, there is limited research on the mechanisms involved, and further in-depth investigation is required.ConclusionThis paper presents a bibliometric analysis of the research framework and hotspots concerning the effect of PM2.5 on ischemic stroke. The analysis aims to provide a comprehensive understanding of this field for researchers. It is expected that research on the effect of PM2.5 on ischemic stroke will remain an important research topic in the future.

Analysis of Predicting Co-Authorship Networks Using Support Vector Machine Model

Analysis of Predicting Co-Authorship Networks Using Support Vector Machine Model

Development and Validation of Machine Learning Algorithms for Prediction of Colorectal Polyps Based on Electronic Health Records.

Colorectal Polyps are the main source of precancerous lesions in colorectal cancer. To increase the early diagnosis of tumors and improve their screening, we aimed to develop a simple and non-invasive diagnostic prediction model for colorectal polyps based on machine learning (ML) and using accessible health examination records. We conducted a single-center observational retrospective study in China. The derivation cohort, consisting of 5426 individuals who underwent colonoscopy screening from January 2021 to January 2024, was separated for training (cohort 1) and validation (cohort 2). The variables considered in this study included demographic data, vital signs, and laboratory results recorded by health examination records. With features selected by univariate analysis and Lasso regression analysis, nine machine learning methods were utilized to develop a colorectal polyp diagnostic model. Several evaluation indexes, including the area under the receiver-operating-characteristic curve (AUC), were used to compare the predictive performance. The SHapley additive explanation method (SHAP) was used to rank the feature importance and explain the final model. 14 independent predictors were identified as the most valuable features to establish the models. The adaptive boosting machine (AdaBoost) model exhibited the best performance among the 9 ML models in cohort 1, with accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and AUC (95% CI) of 0.632 (0.618-0.646), 0.635 (0.550-0.721), 0.674 (0.591-0.758), 0.593 (0.576-0.611), 0.673 (0.654-0.691), 0.608 (0.560-0.655) and 0.687 (0.626-0.749), respectively. The final model gave an AUC of 0.675 in cohort 2. Additionally, the precision recall (PR) curve for the AdaBoost model reached the highest AUPR of 0.648, positioning it nearest to the upper right corner. SHAP analysis provided visualized explanations, reaffirming the critical factors associated with the risk of colorectal polyps in the asymptomatic population. This study integrated the clinical and laboratory indicators with machine learning techniques to establish the predictive model for colorectal polyps, providing non-invasive, cost-effective screening strategies for asymptomatic individuals and guiding decisions for further examination and treatment.

Construction of a predictive model for gastric cancer neuroaggression and clinical validation analysis: A single-center retrospective study.

This study investigated the construction and clinical validation of a predictive model for neuroaggression in patients with gastric cancer. Gastric cancer is one of the most common malignant tumors in the world, and neuroinvasion is the key factor affecting the prognosis of patients. However, there is a lack of systematic analysis on the construction and clinical application of its prediction model. This study adopted a single-center retrospective study method, collected a large amount of clinical data, and applied statistics and machine learning technology to build and verify an effective prediction model for neuroaggression, with a view to providing scientific basis for clinical treatment decisions and improving the treatment effect and survival rate of patients with gastric cancer. To investigate the value of a model based on clinical data, spectral computed tomography (CT) parameters and image omics characteristics for the preoperative prediction of nerve invasion in patients with gastric cancer. A retrospective analysis was performed on 80 gastric cancer patients who underwent preoperative energy spectrum CT at our hospital between January 2022 and August 2023, these patients were divided into a positive group and a negative group according to their pathological results. Clinicopathological data were collected, the energy spectrum parameters of primary gastric cancer lesions were measured, and single factor analysis was performed. A total of 214 image omics features were extracted from two-phase mixed energy images, and the features were screened by single factor analysis and a support vector machine. The variables with statistically significant differences were included in logistic regression analysis to construct a prediction model, and the performance of the model was evaluated using the subject working characteristic curve. There were statistically significant differences in sex, carbohydrate antigen 199 expression, tumor thickness, Lauren classification and Borrmann classification between the two groups (all P < 0.05). Among the energy spectrum parameters, there were statistically significant differences in the single energy values (CT60-CT110 keV) at the arterial stage between the two groups (all P < 0.05) and statistically significant differences in CT values, iodide group values, standardized iodide group values and single energy values except CT80 keV at the portal vein stage between the two groups (all P < 0.05). The support vector machine model with the largest area under the curve was selected by image omics analysis, and its area under the curve, sensitivity, specificity, accuracy, P value and parameters were 0.843, 0.923, 0.714, 0.925, < 0.001, and c:g 2.64:10.56, respectively. Finally, based on the logistic regression algorithm, a clinical model, an energy spectrum CT model, an imaging model, a clinical + energy spectrum model, a clinical + imaging model, an energy spectrum + imaging model, and a clinical + energy spectrum + imaging model were established, among which the clinical + energy spectrum + imaging model had the best efficacy in diagnosing gastric cancer nerve invasion. The area under the curve, optimal threshold, Youden index, sensitivity and specificity were 0.927 (95%CI: 0.850-1.000), 0.879, 0.778, 0.778, and 1.000, respectively. The combined model based on clinical features, spectral CT parameters and imaging data has good value for the preoperative prediction of gastric cancer neuroinvasion.

SOH early prediction of lithium-ion batteries based on voltage interval selection and features fusion

- Accurate state of health (SOH) of batteries is a crucial prerequisite for ensuring the safety and stable operation of electric vehicles. However, existing conventional prediction methods do not consider the influence of early-stage capacity variations on full-cycle SOH with extensive test data. This paper develops a SOH early prediction method of lithium-ion batteries based on voltage interval selection and features fusion. To identify the battery SOH curves with high similarity under identical charge-discharge conditions, a double correlation based early-stage SOH similarity analysis method is presented. To minimize the amount of voltage training data collected during feature extraction, a voltage interval selection method considering sampling time and features correlation is introduced. Meanwhile, a feature fusion method combining entropy weight and correlation factor is used to reduce the impact of redundant health feature information. To address the problems of low accuracy and computational inefficiency using the least squares support vector machine (LSSVM) model, a grey wolf optimizer-LSSVM-adaptive boosting model is developed for battery SOH prediction. The experimental results show that the coefficients of determination remain above 0.98, indicating high SOH prediction accuracy using the developed method. Compared to other methods, the mean absolute errors based on the developed method are maintained below 1.5 %.

Prediction of Rock Unloading Strength Based on PSO-XGBoost Hybrid Models.

Rock excavation is essentially an unloading behavior, and its mechanical properties are significantly different from those under loading conditions. In response to the current deficiencies in the peak strength prediction of rocks under unloading conditions, this study proposes a hybrid learning model for the intelligent prediction of the unloading strength of rocks using simple parameters in rock unloading tests. The XGBoost technique was used to construct a model, and the PSO-XGBoost hybrid model was developed by employing particle swarm optimization (PSO) to refine the XGBoost parameters for better prediction. In order to verify the validity and accuracy of the proposed hybrid model, 134 rock sample sets containing various common rock types in rock excavation were collected from international and Chinese publications for the purpose of modeling, and the rock unloading strength prediction results were compared with those obtained by the Random Forest (RF) model, the Support Vector Machine (SVM) model, the XGBoost (XGBoost) model, and the Grid Search Method-based XGBoost (GS-XGBoost) model. Meanwhile, five statistical indicators, including the coefficient of determination (R2), mean absolute error (MAE), mean absolute percentage error (MAPE), mean square error (MSE), and root mean square error (RMSE), were calculated to check the acceptability of these models from a quantitative perspective. A review of the comparison results revealed that the proposed PSO-XGBoost hybrid model provides a better performance than the others in predicting rock unloading strength. Finally, the importance of the effect of each input feature on the generalization performance of the hybrid model was assessed. The insights garnered from this research offer a substantial reference for tunnel excavation design and other representative projects.

A stacking ensemble machine learning model for evaluating cardiac toxicity of drugs based on in silico biomarkers.

This study addresses the critical issue of drug-induced torsades de pointes (TdP) risk assessment, a vital aspect of new drug development due to its association with arrhythmia and sudden cardiac death. Existing methodologies, particularly those reliant on a single biomarker derived from CiPA O'Hara-Rudy (CiPAORdv1.0) ventricular cell model without the hERG dynamic as input to the individual machine learning model, have limitations in capturing the complexity inherent in the comprehensive range of factors influencing drug-induced TdP risk. This study aims to overcome these limitations by proposing a stacking ensemble machine learning approach by integrating multiple in silico biomarkers derived from the CiPAORdv1.0 with hERG dynamic characteristics. The ensemble machine learning model consisted of three artificial neural network (ANN) models as baseline model and support vector machine (SVM), logistic regression (LR), random forest (RF), and extreme gradient boosting (XGBoost) models as meta-classifier. The highest AUC score of 1.00 (0.90-1.00) for high risk, 0.97 (0.84-1.00) for intermediate risk, and 1.00 (0.87-1.00) for low risk were obtained using seven biomarkers derived from the CiPAORdv1.0 with hERG dynamic characteristics. Furthering our investigation, we explored the model's robustness by incorporating interindividual variability into the generation of in silico biomarkers from a population of human ventricular cell models. This study also enabled an analysis of TdP risk classification under high clinical exposure and therapeutic scenarios for several drugs. Additionally, from a sensitivity analysis, we revealed four important ion channels, namely, CaL, NaL, Na, and Kr channels that affect significantly the important biomarkers for TdP risk prediction.