Search Results (728)

This paper presents an application of the Ant Colony Optimization (ACO) algorithm combined with the Logistic Regression (LR) method in the lead acid battery charging process. The ACO algorithm is used to obtain the best current pattern in the battery charging system to produce a smart charging system with a fast and safe charging current for the battery. The best current pattern is conducted gradually and repeatedly to obtain termination in the form of the best current pattern according to the ACO algorithm. The results of the algorithm design produce a current pattern consisting of 10 A, 5 A, 3 A, 2 A, and 0 A. The charging system with this algorithm can charge all types of lead acid batteries. In this research, the capacity of battery 1’s State of Charge (SOC) is 56%, battery 2’s SOC is 62%, and battery 3’s SOC is 80%. When recharging the battery’s full condition to a SOC of 100%, the length of time for charging battery 1 for 12.73 min, battery 2 takes 15.73 min, and battery 3 takes 29.11 min. Smart charging with the ACO can charge the battery safely without current fluctuations compared to charging without an algorithm such that the amount of charging current used is not dangerous for the battery. In addition, data analysis is carried out to determine the value of accuracy in estimating SOC charging using supervised learning linear regression. The results of the data analysis with linear regression show that the battery’s SOC estimation has good accuracy, with an RMSE value of 0.32 and an MAE of 0.27. Full article

Complex diseases pose challenges in prediction due to their multifactorial and polygenic nature. This study employed machine learning (ML) to analyze genomic data from the UK Biobank, aiming to predict the genomic predisposition to complex diseases like multiple sclerosis (MS) and Alzheimer’s disease (AD). We tested logistic regression (LR), ensemble tree methods, and deep learning models for this purpose. LR displayed remarkable stability across various subsets of data, outshining deep learning approaches, which showed greater variability in performance. Additionally, ML methods demonstrated an ability to maintain optimal performance despite correlated genomic features due to linkage disequilibrium. When comparing the performance of polygenic risk score (PRS) with ML methods, PRS consistently performed at an average level. By employing explainability tools in the ML models of MS, we found that the results confirmed the polygenicity of this disease. The highest-prioritized genomic variants in MS were identified as expression or splicing quantitative trait loci located in non-coding regions within or near genes associated with the immune response, with a prevalence of human leukocyte antigen (HLA) gene annotations. Our findings shed light on both the potential and the challenges of employing ML to capture complex genomic patterns, paving the way for improved predictive models. Full article

Background: Cardiovascular disease (CVD) is a prominent determinant of mortality, accounting for 17 million lives lost across the globe each year. This underscores its severity as a critical health issue. Extensive research has been undertaken to refine the forecasting of CVD in patients using various supervised, unsupervised, and deep learning approaches. Methods: This study presents HeartEnsembleNet, a novel hybrid ensemble learning model that integrates multiple machine learning (ML) classifiers for CVD risk assessment. The model is evaluated against six classical ML classifiers, including support vector machine (SVM), gradient boosting (GB), decision tree (DT), logistic regression (LR), k-nearest neighbor (KNN), and random forest (RF). Additionally, we compare HeartEnsembleNet with Hybrid Random Forest Linear Models (HRFLM) and ensemble techniques including stacking and voting. Results: Employing a dataset of 70,000 cardiac patients with 12 clinical attributes, our proposed model achieves a notable accuracy of 92.95% and a precision of 93.08%. Conclusions: These results highlight the effectiveness of hybrid ensemble learning in enhancing CVD risk prediction, offering a promising framework for clinical decision support. Full article

Background: The rate of morbidity after liver surgery is estimated at 30% and can be even higher when considering higher-risk subgroups of patients. Frailty is believed to better predict surgical outcomes by showcasing the patient’s ability to withstand major surgical stress and selecting frail ones. Methods: This is a single-centre, observational case–control study on patients diagnosed with liver malignancies who underwent liver resections between 2013 and 2024. The five-item modified Frailty Index (mFI-5) was used to split patients into frail and non-frail. The two groups were compared in terms of preoperative, operative and postoperative outcomes using a chi-squared and logistic regression model. Results: A total of 230 patients were included and split into two groups: non-frail, NF, n = 90, and frail patients, F, n = 140. Overall, F patients had a higher rate of morbidity (p = 0.04) but with similar mortality and length of stay. When considering only major liver resections, F patients had a higher probability of posthepatectomy liver failure (LR 6.793, p = 0.009), postoperative bleeding (LR 9.541, p = 0.002) and longer ICU stay (LR 8.666, p = 0.003), with similar rates of bile leak, surgical site infections, length of stay and mortality. Conclusions: Frailty seems to be a solid predictor of posthepatectomy liver failure in patients undergoing major liver resections and is associated with a longer ICU stay. However, mortality and surgical morbidity seem to be comparable between frail and non-frail patients. Full article

Background: Surgical site infections (SSIs) lead to higher hospital readmission rates and healthcare costs, representing a significant global healthcare burden. Machine learning (ML) has demonstrated potential in predicting SSIs; however, the challenge of addressing imbalanced class ratios remains. Objectives: The aim of this study is to evaluate and enhance the predictive capabilities of machine learning models for SSIs by assessing the effects of feature selection, resampling techniques, and hyperparameter optimization. Methods: Using routine SSI surveillance data from multiple hospitals in Saudi Arabia, we analyzed a dataset of 64,793 surgical patients, of whom 1632 developed SSI. Seven machine learning algorithms were created and tested: Decision Tree (DT), Gaussian Naive Bayes (GNB), Support Vector Machine (SVM), Logistic Regression (LR), Random Forest (RF), Stochastic Gradient Boosting (SGB), and K-Nearest Neighbors (KNN). We also improved several resampling strategies, such as undersampling and oversampling. Grid search five-fold cross-validation was employed for comprehensive hyperparameter optimization, in conjunction with balanced sampling techniques. Features were selected using a filter method based on their relationships with the target variable. Results: Our findings revealed that RF achieves the highest performance, with an MCC of 0.72. The synthetic minority oversampling technique (SMOTE) is the best-performing resampling technique, consistently enhancing the performance of most machine learning models, except for LR and GNB. LR struggles with class imbalance due to its linear assumptions and bias toward the majority class, while GNB’s reliance on feature independence and Gaussian distribution make it unreliable for under-represented minority classes. For computational efficiency, the Instance Hardness Threshold (IHT) offers a viable alternative undersampling technique, though it may compromise performance to some extent. Conclusions: This study underscores the potential of ML models as effective tools for assessing SSI risk, warranting further clinical exploration to improve patient outcomes. By employing advanced ML techniques and robust validation methods, these models demonstrate promising accuracy and reliability in predicting SSI events, even in the face of significant class imbalances. In addition, using MCC in this study ensures a more reliable and robust evaluation of the model’s predictive performance, particularly in the presence of an imbalanced dataset, where other metrics may fail to provide an accurate evaluation. Full article

This study evaluates the effectiveness of six machine learning models, Artificial Neural Networks (ANN), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Logistic Regression (LR), for predictive maintenance in building systems. Utilizing a high-resolution dataset collected every five minutes from six office rooms at Aalborg University in Denmark over a ten-month period (27 February 2023 to 31 December 2023), we defined rule-based conditions to label historical faults in HVAC, lighting, and occupancy systems, resulting in over 100,000 fault instances. XGBoost outperformed other models, achieving an accuracy of 95%, precision of 93%, recall of 94%, and an F1-score of 0.93, with a computation time of 60 s. The model effectively predicted critical faults such as “Light_On_No_Occupancy” (1149 occurrences) and “Damper_Open_No_Occupancy” (8818 occurrences), demonstrating its potential for real-time fault detection and energy optimization in building management systems. Our findings suggest that implementing XGBoost in predictive maintenance frameworks can significantly enhance fault detection accuracy, reduce energy waste, and improve operational efficiency. Full article

Background: Pressure injuries (PIs) are increasing worldwide, and there has been no significant improvement in preventing them. Traditional assessment tools are widely used to identify a patient at risk of developing a PI. This study aims to construct a novel fused multi-channel prediction model of PIs in adult hospitalized patients using machine learning algorithms (MLAs). Methods: A multi-phase quantitative approach involving a case–control experimental design was used. A first-hand dataset was collected retrospectively between March/2022 and August/2023 from the electronic medical records of three hospitals in Palestine. Results: The total number of patients was 49,500. A balanced dataset was utilized with a total number of 1110 patients (80% training and 20% testing). The models that were developed utilized eight MLAs, including linear regression and support vector regression (SVR), logistic regression (LR), random forest (RF), gradient boosting (GB), K-nearest neighbor (KNN), decision tree (DT), and extreme gradient boosting (XG boosting) and validated with five-fold cross-validation techniques. The best model was RF, for which the accuracy was 0.962, precision was 0.942, recall was 0.922, F1 was 0.931, area under curve (AUC) was 0.922, false positive rate (FPR) was 0.155, and true positive rate (TPR) was 0.782. Conclusions: The predictive factors were age, moisture, activity, length of stay (LOS), systolic blood pressure (BP), and albumin. A novel fused multi-channel prediction model of pressure injury was developed from different datasets. Full article

Background: The implementation of machine learning (ML) models has significantly impacted neuroimaging. Recent data suggest that these models may improve the accuracy of diagnosing and predicting outcomes in patients with Chiari malformation type I (CMI). Methods: A scoping review was conducted according to the guidelines put forth by PRISMA. The literature search was performed in PubMed/MEDLINE, SCOPUS, and ScienceDirect databases. We included observational or experimental studies focusing on the applications of ML in patients with CMI. Results: A total of 9 articles were included. All the included articles were retrospective. Five out of the nine studies investigated the applicability of machine learning models for diagnosing CMI, whereas the remaining studies focused on the prognosis of the patients treated for CM. Overall, the accuracy of the machine learning models utilized for the diagnosis ranged from 0.555 to 1.00, whereas the specificity and sensitivity ranged from 0.714 to 1.00 and 0.690 to 1.00, respectively. The accuracy of the prognostic ML models ranged from 0.402 to 0.820, and the AUC ranged from 0.340 to 0.990. The most utilized ML model for the diagnosis of CMI is logistic regression (LR), whereas the support vector machine (SVM) is the most utilized model for postoperative prognosis. Conclusions: In the present review, both conventional and novel ML models were utilized to diagnose CMI or predict patient outcomes following surgical treatment. While these models demonstrated significant potential, none were highly validated. Therefore, further research and validation are required before their actual implementation in standard medical practice. Full article

The grade of Daqu significantly influences the quality of Baijiu. To address the issues of high subjectivity, substantial labor costs, and low detection efficiency in Daqu grade evaluation, this study focused on light-flavor Daqu and proposed a two-layer classification structure model based on computer vision and machine learning. Target images were extracted using three image segmentation methods: threshold segmentation, morphological fusion, and K-means clustering. Feature factors were selected through methods including mean decrease accuracy based on random forest (RF-MDA), recursive feature elimination (RFE), LASSO regression, and ridge regression. The Daqu grade evaluation model was constructed using support vector machine (SVM), logistic regression (LR), random forest (RF), k-nearest neighbor (KNN), and a stacking model. The results indicated the following: (1) In terms of image segmentation performance, the morphological fusion method achieved an accuracy, precision, recall, F1-score, and AUC of 96.67%, 95.00%, 95.00%, 0.95, and 0.96, respectively. (2) For the classification of Daqu-P, Daqu-F, and Daqu-S, RF models performed best, achieving an accuracy, precision, recall, F1-score, and AUC of 96.67%, 97.50%, 97.50%, 0.97, and 0.99, respectively. (3) In distinguishing Daqu-P from Daqu-F, the combination of the RF-MDA method and the stacking model demonstrated the best performance, with an accuracy, precision, recall, F1-score, and AUC of 90.00%, 94.44%, 85.00%, 0.89, and 0.95, respectively. This study provides theoretical and technical support for efficient and objective Daqu grade evaluation. Full article

Background/Objectives: The methacholine bronchial provocation test (MBPT) is a diagnostic test frequently used to evaluate airway hyper-reactivity. MBPT is essential for diagnosing asthma; however, it can be time-consuming and resource-intensive. This study aimed to develop an artificial intelligence (AI) model to predict the MBPT results using forced expiratory volume in one second (FEV₁) and bronchodilator test measurements from spirometry. Methods: a dataset of spirometry measurements, including Pre- and Post-bronchodilator FEV₁, was used to train and validate the model. Results: Among the evaluated models, the multilayer perceptron (MLP) achieved the highest area under the curve (AUC) of 0.701 (95% CI: 0.676–0.725), accuracy of 0.758, and an F1-score of 0.853. Logistic regression (LR) and a support vector machine (SVM) demonstrated comparable performance with AUC values of 0.688, while random forest (RF) and extreme gradient boost (XGBoost) achieved slightly lower AUC values of 0.669 and 0.672, respectively. Feature importance analysis of the MLP model identified key contributing features, including Pre-FEF_25–75 (%), Pre-FVC (L), Post FEV₁/FVC, Change-FEV₁ (L), and Change-FEF_25–75 (%), providing insight into the interpretability and clinical applicability of the model. Conclusions: These results highlight the potential of the model to utilize readily available spirometry data, particularly FEV₁ and bronchodilator responses, to accurately predict MBPT results. Our findings suggest that AI-based prediction can improve asthma diagnostic workflows by minimizing the reliance on MBPT and enabling faster and more accessible assessments. Full article

In the digital era, cybersecurity is integral for preserving national security, digital privacy, and social sustainability. This research emphasizes the role of non-functional equirements (NFRs) in developing secure software systems that enhance societal wellbeing by ensuring data protection, user privacy, and system robustness. Specifically, this study introduces a proof-of-concept approach by leveraging machine learning (ML) models to classify NFRs and identify security-related issues early in the software development lifecycle. Two experiments were conducted to assess the effectiveness of different models for binary and multi-class classification tasks. In Experiment 1, BERT-based models and artificial neural networks (ANNs) were fine-tuned to classify NFRs into security and non-security categories using a dataset of 803 statements. BERT-based models outperformed ANNs, achieving higher accuracy, precision, recall, and ROC-AUC scores, with hyperparameter tuning further enhancing the results. Experiment 2 assessed logistic regression (LR), a support vector machine (SVM), and XGBoost for the multi-class classification of security-related NFRs into seven categories. The SVM and XGBoost showed strong performance, achieving high precision and recall in specific categories. The findings demonstrate the effectiveness of advanced ML models in automating NFR classification, improving software security, and supporting social sustainability. Future work will explore hybrid approaches to enhance scalability and accuracy. Full article

Introduction: The severity of carpal tunnel syndrome (CTS) is evaluated by electrophysiological examination as well as a patient-oriented questionnaire. We hypothesized that machine learning could predict postoperative electrophysiological severity as well as the scores of patient-oriented questionnaires. In this study, we developed machine learning models to predict postoperative changes in electrophysiological severity and changes in the Carpal Tunnel Syndrome Instrument (CTSI). Materials and Methods: Data from four hundred and twenty hands of individuals who had been diagnosed with CTS and undergone carpal tunnel release were used. The features used for the machine learning model were preoperative age, gender, distal motor latency (DML) value, sensory nerve conduction velocity (SCV) value, preoperative electrophysiological severity stage, CTSI-SS value, and CTSI-FS value. Logistic Regression (LR), ElesticNet (EN), Support Vector Machine (SVM), Random Forest (RF), and LightGBM (LGBM) were used as machine learning algorithms. A machine learning model was created to binary classify the electrophysiologic severity at one year postoperatively. In the second experiment, regression models were created to predict the change in CTSI-SS and CTSI-FS at one year postoperatively. Results: In the electrophysiological severity classification model, LGBM showed the highest score (AUC = 0.802). Preoperative DML, age, and preoperative electrophysiological severity were important factors for model prediction. RF model showed the best performance. In the regression model predicting the change in CTSI-SS or CTSI-FS (RMSE: 0.418, 0.333, respectively), preoperative age and CTSI-SS or CTSI-FS scores were important factors for model prediction. Conclusions: The machine learning model can predict postoperative electrophysiological severity and CTSI score with high accuracy. Full article

Background/Objective: Subarachnoid hemorrhage (SAH) is associated with high morbidity and mortality rates, necessitating prognostic algorithms to guide decisions. Our study evaluates the use of machine learning (ML) models for predicting 1-month and 1-year mortality among SAH patients using national electronic health records (EHR) system. Methods: Retrospective cohort of 29,274 SAH patients, identified through national EHR system from January 2017 to December 2022, was analyzed, with mortality data obtained from central civil registration system in Türkiye. Variables included (n = 102) pre- (n = 65) and post-admission (n = 37) data, such as patient demographics, clinical presentation, comorbidities, laboratory results, and complications. We employed logistic regression (LR), decision trees (DTs), random forests (RFs), and artificial neural networks (ANN). Model performance was evaluated using area under the curve (AUC), average precision, and accuracy. Feature significance analysis was conducted using LR. Results: The average age was 56.23 ± 16.45 years (47.8% female). The overall mortality rate was 22.8% at 1 month and 33.3% at 1 year. One-month mortality increased from 20.9% to 24.57% (p < 0.001), and 1-year mortality rose from 30.85% to 35.55% (p < 0.001) in the post-COVID period compared to the pre-COVID period. For 1-month mortality prediction, the ANN, LR, RF, and DT models achieved AUCs of 0.946, 0.942, 0.931, and 0.916, with accuracies of 0.905, 0.901, 0.893, and 0.885, respectively. For 1-year mortality, the AUCs were 0.941, 0.927, 0.926, and 0.907, with accuracies of 0.884, 0.875, 0.861, and 0.851, respectively. Key predictors of mortality included age, cardiopulmonary arrest, abnormal laboratory results (such as abnormal glucose and lactate levels) at presentation, and pre-existing comorbidities. Incorporating post-admission features (n = 37) alongside pre-admission features (n = 65) improved model performance for both 1-month and 1-year mortality predictions, with average AUC improvements of 0.093 ± 0.011 and 0.089 ± 0.012, respectively. Conclusions: Our study demonstrates the effectiveness of ML models in predicting mortality in SAH patients using big data. LR models’ robustness, interpretability, and feature significance analysis validate its importance. Including post-admission data significantly improved all models’ performances. Our results demonstrate the utility of big data analytics in population-level health outcomes studies. Full article

As pension benefits from statutory public schemes become less generous, and many countries face pension-savings crises, the willingness to participate in supplementary retirement saving instruments becomes crucial for sustainable financial well-being. The main objective of this article is to present how trust and financial literacy influence the choice of sustainable auto-enrolment pension scheme as a private and supplementary pension savings. The study highlighted factors influencing participation in auto-enrollment and private supplementary pension savings. The study focuses mainly on financial literacy and trust. We used the CAWI method with 857 interviews in Poland—the first country in Central and Eastern Europe to introduce an auto-enrolment pension system. Our study uses multivariable data-mining tools, and several regression models were applied. We used Logistic Regression (LR), Multivariate Linear Regression (MLR), and Factor Analysis of Mixed Data (FAMD) to support the LR analysis. We propose four regression models. Our findings present that: 1. The lower the consumer’s knowledge level, the more their decisions are based on trust. 2. Trust in the state, rather than trust in financial institutions, plays a crucial role for people with low financial literacy, which is a critical factor in choosing the auto-enrolment option for pension savings. 3. Men had higher odds of auto-enrolment pension saving than women. 4. Employees of economic universities and academics had higher odds of participating in capital pension plans than those of general universities and non-academics. Our findings can signal to governments and policymakers about factors influencing the choice of auto-enrolment supplementary retirement savings. These findings strengthen the role of sustainable economic education. Full article

Critical National Infrastructure includes large networks such as telecommunications, transportation, health services, police, nuclear power plants, and utilities like clean water, gas, and electricity. The protection of these infrastructures is crucial, as nations depend on their operation and stability. However, cyberattacks on such systems appear to be increasing in both frequency and severity. Various machine learning approaches have been employed for anomaly detection in Critical National Infrastructure, given their success in identifying both known and unknown attacks with high accuracy. Nevertheless, these systems are vulnerable to adversarial attacks. Hackers can manipulate the system and deceive the models, causing them to misclassify malicious events as benign, and vice versa. This paper evaluates the robustness of traditional machine learning techniques, such as Support Vector Machines (SVMs) and Logistic Regression (LR), as well as Artificial Neural Network (ANN) algorithms against adversarial attacks, using a novel dataset captured from a model of a clean water treatment system. Our methodology includes four attack categories: random label flipping, targeted label flipping, the Fast Gradient Sign Method (FGSM), and Jacobian-based Saliency Map Attack (JSMA). Our results show that, while some machine learning algorithms are more robust to adversarial attacks than others, a hacker can manipulate the dataset using these attack categories to disturb the machine learning-based anomaly detection system, allowing the attack to evade detection. Full article