Search Results (1,881)

Background/Objectives: Motor abnormalities are frequently observed in schizophrenia, but accessible methods for objective gait quantification remain limited. This exploratory controlled-setting study examined whether markerless smartphone-based video analysis combined with interpretable machine learning could quantify gait-related motor phenotypes in individuals with schizophrenia. Methods: Gait videos were collected from 100 individuals with schizophrenia and 35 healthy controls using a single-site, single-device, standardized recording setup. MediaPipe Pose was used to extract skeletal landmarks and derive 12 image-plane spatiotemporal and estimated two-dimensional knee-kinematic gait features. After temporal segmentation and quality control, 404 usable gait segments derived from 135 participants were analyzed as repeated segment-level observations. Decision Tree and Support Vector Machine models were applied for exploratory segment-level group-separation analysis using segment-wise 15-fold cross-validation after the full post-quality-control dataset had been balanced before fold allocation. Results: Several extracted gait features differed between groups, particularly image-plane ankle displacement, mean step displacement, displacement velocity, step characteristics, and knee-joint motion. In the Decision Tree model, image-plane ankle displacement served as the primary root node, indicating its central role in internal segment-level group separation. However, the healthy control group was substantially younger and not age-matched. In addition, the segment-level statistical comparisons did not account for within-participant clustering. Accordingly, the reported p values and confidence intervals may overstate statistical precision. Separately, segment-wise cross-validation allowed segments from the same participant to occur across folds and resampling was performed before fold partitioning. Because oversampling was performed with replacement, duplicated segment instances could also occur across training and validation folds. Consequently, the statistical findings should be interpreted as exploratory segment-level patterns rather than participant-level inference, and the machine-learning performance estimates should be regarded only as potentially optimistic apparent internal segment-level results and should not be regarded as evidence of participant-level generalization, diagnostic validity, screening accuracy, or clinical applicability. Conclusions: Markerless video-based gait analysis with interpretable machine learning may provide a feasible research-support approach for quantifying gait-related motor phenotypes in individuals with schizophrenia. These findings should not be interpreted as evidence for participant-level clinical classification, diagnostic or screening validity, clinical utility, or readiness for deployment, or as proof of cross-device or cross-environment reproducibility. Future studies require matched controls, psychiatric comparison groups, subject-wise validation, external datasets, calibrated gait measures, systematic cross-configuration reproducibility testing, and privacy-preserving data governance. Full article

Background: Falls are the most common hospital-acquired adverse event, leading to extended hospitalization, loss of independence, disability, and premature death. Routine fall risk assessments are time-consuming, even with limited factors. An AI-derived fall prediction model can provide more comprehensive and comparably accurate risk predictions quickly and as often as needed. Objective: To develop and validate a fall prediction model for fall risk in adult inpatients. Methods: Patient records from 2016 were extracted from the adult inpatient database, including information from the Electronic Inpatient Medication Records, SAP, and Hospital Incident Reporting System. The sample consisted of 1506 cases (1:5 faller to non-faller). The fall prediction model was trained using the following four variables: demographics, diagnosis, medications, and surgery. Data sources included the hospital’s data repository, integrating admission/discharge, pharmacy, laboratory, and incident reports. Results: The support vector machine model performed best among all tested models, achieving an AUC of 0.803, recall of 0.816, and precision of 0.440. In the validation cohort (978 patients: 163 fallers and 815 non-fallers), the fall prediction model demonstrated moderate-to-good discrimination (AUC 0.79), with accuracy of 0.67, sensitivity of 0.46, and specificity of 0.86. Compared with the nursing four-item fall risk assessment, which showed lower discrimination (AUC 0.65, accuracy 0.65, sensitivity 0.58, specificity 0.72), the fall prediction model had better specificity and overall discrimination, though the nursing tool was more sensitive in identifying fallers. Conclusions: The fall prediction model using demographics, diagnoses, medication, and surgery data predicts falls risk effectively. It enables timely, accurate risk assessments and supports preventive interventions, saving nurses’ time for direct patient care. Full article

The increasing frequency and growing impact of cyberattacks have led organizations to adopt proactive defense approaches to cybersecurity risk mitigation, especially in the case of advanced persistent threats (APTs). The correct identification of the specific malicious actors behind a cyberattack is important for the success of incident response and for the investigative work of the security operations center (SOC) team. This research explores the capabilities and limitations of a machine learning (ML) approach to identifying malicious actors and the threats they pose (threat actor attribution) based on the tactics, techniques, and procedures (TTP) observed in specific cybersecurity incidents and on the incident context (the geographical location and industry affiliation of the victims targeted in the attack). A large language model (LLM) was used to extract TTPs from the MITRE ATT&CK database of cybersecurity incidents. The experiments included modeling threat actor attribution using five ML algorithms: k-nearest neighbors (KNN), decision tree (DT), random forest (RF), support vector machine (SVM), and naïve Bayes (NB), with different methods applied for feature selection and weighting. The results indicated that model accuracy and other performance metrics were significantly improved when the input dataset included both TTP and contextual features. The KNN and SVM models produced the best performance results; the highest classification accuracy achieved was 93.19%. The outcomes of this study may be applied by cybersecurity professionals to identify malicious actors, estimate the number and types of data points that are required to adequately attribute a cyberattack to an actor, and improve the accuracy of the classification by weighting the input dataset features. Full article

Introduction: Voice-based digital biomarkers have emerged as a promising approach for distinguishing individuals with neurodegenerative disorders, particularly Parkinson’s disease (PD), from healthy subjects (HS). With the increasing availability of smartphone and web-based recording tools, voice data can be collected efficiently in both clinical and remote settings. However, further validation is required before such approaches can be translated into routine clinical practice. Methods: This study used a cross-sectional analysis at the recording level, treating repeated recordings from the same participant as separate observations collected at Vito Fazzi Hospital in Lecce, Italy. Speech recordings from individuals with Parkinson’s disease (PD) and healthy controls were collected using the dedicated Talia smartphone and web application. Sustained vowel phonation (/a/) was analyzed as the primary speech task. Following data acquisition, feature extraction was performed as a crucial step in the speech analysis pipeline, as the quality and relevance of the extracted features directly influence the ability of machine learning models to discriminate between Parkinson’s disease (PD) patients and healthy controls. To capture various aspects of speech impairment associated with PD, a comprehensive set of acoustic features was extracted, including long-term features (pitch, jitter, and shimmer), nonlinear descriptors such as Recurrence Period Density Entropy (RPDE), and short-term feature based on Mel-Frequency Cepstral Coefficients (MFCCs). These features were subsequently used to develop and evaluate machine learning models for the classification of Parkinson’s disease and healthy subjects. Feature selection was performed using SHAP to identify the most informative vocal biomarkers. Model performance was assessed using five independent random train–test splits (70% training and 30% testing), supported by an internal five-fold cross-validation procedure within the training data. Multiple machine learning models were developed and evaluated, including Random Forest, Logistic Regression, Support Vector Machine, Naive Bayes, K-Nearest Neighbors, Decision Tree, Artificial Neural Network, and Gradient Boosting. Results: The evaluated models demonstrated strong recording-level classification performance. Artificial Neural Networks (ANN) and K-Nearest Neighbors (KNN) achieved the highest accuracy scores (0.9545 and 0.9494, respectively), along with superior recall (up to 0.9500), precision (up to 0.9551), and F1-score (up to 0.9525). Both models also exhibited excellent discriminative ability, with ROC-AUC values reaching 0.9882 (ANN) and 0.9893 (KNN). In contrast, Naive Bayes and Decision Tree showed comparatively lower performance across all metrics. Log-loss analysis further confirmed the robustness of ANN and KNN, which achieved the lowest values (0.2552 and 0.2510, respectively), indicating well-calibrated predictions. Overall, the findings highlight the consistency and generalizability of ANN and KNN across cross-validation splits. Conclusions: This study demonstrates that machine learning models, particularly ANN and KNN, can effectively differentiate Parkinson’s disease from healthy conditions using voice recordings. The integration of explainable AI for feature selection enhances model transparency and clinical relevance. However, the reported performance estimates were obtained from a recording-level analysis and should be interpreted as preliminary findings. Further studies involving larger cohorts and participant-level validation strategies are required to determine the generalizability and clinical applicability of these approaches. Full article

Coal gangue concrete (CGC) is an effective green building material that can promote the resource utilization of solid waste. To study its durability performance and degradation mechanism under sulfate attack with dry-wet cycles, and to realize the intelligent prediction of mechanical properties, this study prepared CGC specimens with a water-to-binder ratio of 0.4, a fine aggregate replacement rate of 20%, and coarse aggregate replacement rates of 0%, 20%, 50%, 80%, and 100%. The specimens were tested under 30, 60, 90, and 120 dry-wet cycles in 10% MgSO₄ solution. Mass loss, relative dynamic elastic modulus, and compressive and flexural strength corrosion resistance coefficients were used as evaluation indices, and SEM and XRD were adopted to analyze microstructural deterioration. A database compiled from literature data was established, and six machine learning models-random forest (RF), artificial neural network (ANN), decision tree (DT), support vector machine (SVM), particle swarm optimization-artificial neural network (PSO-ANN), and particle swarm optimization-support vector machine (PSO-SVM) were constructed to predict the strength corrosion resistance coefficients. Test results indicate that all macroscopic indices first increased and then decreased with the number of dry-wet cycles. Early ettringite and gypsum products filled internal pores, while prolonged sulfate attack caused decalcification and structural degradation of C-S-H gel, resulting in obvious performance loss. The PSO-SVM model showed the best prediction accuracy, with R² values of 0.912 and 0.981 for compressive and flexural strength corrosion resistance coefficients, respectively. Feature importance analysis shows that dry-wet cycles had the most significant negative impact, followed by the coal gangue fine aggregate replacement rate. This study provides support for the durability evaluation and intelligent prediction of coal gangue concrete in sulfate environments. Full article

The side ledge is essential for maintaining thermal stability and protecting the sidewall lining in an aluminum reduction cell. Its shape and temperature distribution strongly influence cell operation and energy efficiency. However, conventional computational fluid dynamics (CFD) methods, while accurate, are computationally intensive and unsuitable for fast evaluation under changing operating conditions. To overcome this limitation, this study develops a CFD-driven machine learning surrogate model for rapid prediction of cross-sectional temperature fields and side ledge morphology in an aluminum reduction cell under magnetohydrodynamic (MHD) conditions. High-fidelity training data are first generated from MHD-CFD simulations. Key physical variables, including heat generation, effective thermal transport properties, and velocity components, are extracted from multiple cross-sections prior to side ledge formation to form the input dataset. Five machine learning methods—Decision Tree, Random Forest, XGBoost, K-Nearest Neighbors, and Support Vector Regression—are then developed to predict the resulting temperature field and side ledge profile after solidification. The performance of each model is systematically compared in terms of accuracy and applicability. Among the tested models, Random Forest demonstrates the best overall performance. Its predicted temperature fields closely match CFD results, with a maximum temperature deviation below 0.23 °C, while the average side ledge thickness error is only 0.0153 m. Moreover, the proposed surrogate model reduces computation time from approximately 168 h to 3 min compared with CFD simulations, while maintaining high predictive accuracy. Overall, the proposed method enables fast and accurate evaluation of thermal states, supports operational optimization, and provides a foundation for digital twin development of aluminum reduction cells. Full article

The evolution of Advanced Metering Infrastructure (AMI) requires reliable, energy-efficient, and scalable communication technologies for connecting large numbers of smart meters and gateways with utility backend systems. Among 3GPP Cellular Internet of Things (CIoT) technologies, Narrowband IoT (NB-IoT) and LTE-M are promising candidates due to their extended coverage, low cost, and power efficiency. However, selecting between them remains challenging because performance depends on deployment environments, spatial distribution, and radio signal conditions. This study addresses the CIoT network selection problem in AMI networks by applying machine learning to predict the appropriate communication technology from smart meter location and Reference Signal Received Power (RSRP). Three supervised learning algorithms, namely Decision Tree, Support Vector Machine, and XGBoost, were evaluated using field measurement datasets from two AMI deployment areas. A spatial holdout strategy was applied to assess performance in unseen geographical regions. Decision Tree achieved the best performance in Area 1, with an accuracy of 0.7143 and an F1-score of 0.6154. In Area 2, XGBoost achieved the highest performance, with an accuracy of 0.9732 and an F1-score of 0.9388. The results demonstrate the feasibility of ML-based CIoT selection under spatially heterogeneous and imbalanced deployment conditions. Full article

In this article, machine learning (ML) is proposed as a solution to detect and classify false message injection attacks in autonomous ground navigation. First, multiple trajectories are designed and simulated to collect authentic feature samples offered by the odometry and inertial measurement unit (IMU) of an autonomous ground vehicle (UGV). Then, a dataset comprising these samples and other injected samples that simulate two cyberattacks, namely path modification (PM) and velocity drift (VD), is created to train, validate, and benchmark various ML classification models. These include decision tree (DT), k-nearest neighbors (KNN), multi-layer perceptron (MLP), random forest (RF), and support vector machine (SVM). The optimum classification model is experimentally evaluated using a UGV platform, and results suggest that the proposed solution allows the detection of authentic and attacked messages with more than 98% average accuracy and sub-millisecond prediction time. Thus, this solution is ideal for real-time classification, especially in fixed-route applications, e.g., public transportation. Full article

Background/Objectives: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder necessitating accurate and timely diagnosis for effective clinical intervention. While deep learning methods have shown promise in AD classification, many rely on computationally intensive architectures and high-dimensional feature representations. This study introduces a lightweight hybrid framework combining deep feature extraction, dimensionality reduction, and adaptive classification for MRI-based Alzheimer’s disease stage classification. Methods: Utilizing MRI images from a publicly available Alzheimer’s disease dataset encompassing four clinical stages (Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented), deep features were extracted using a pre-trained SqueezeNet model as a fixed feature extractor, generating 1000-dimensional feature vectors. Due to the computational complexity and for the improvement of the model efficiency, the dimensionality reduction technique, Principal Component Analysis (PCA) was then applied. This resulted in an optimum representation of 100 principal components, retaining about 96% of the variance. Then, the performances of various machine learning classifiers such as k-Nearest Neighbors (kNN), Support Vector Machine (SVM), Decision Tree (DT), Neural Network (NN), Naïve Bayes (NB), Logistic Regression (LR) and Enhanced Fuzzy Min–Max Neural Network (EFMM) were tested. The accuracy, precision, recall, F1-score, area under the receiver operating characteristic curve (AUC), and confusion matrices were used to evaluate the performance. Stratified 5-fold cross validation was used to ensure the strength of our results. Results: The findings show that PCA has a significant improvement in classification accuracy for most of the models. In particular, the EFMM classifier outperformed the other classifiers, with an accuracy of 97.19% on the independent test set. After PCA, the AUC values for classes such as Mild Demented, Moderate Demented, Non-Demented and Very Mild Demented were obtained as 97.12%, 99.97%, 93.79% and 95.26% respectively. We further validated our proposed framework using stratified 5-fold cross validation which further corroborated the robustness of our proposed framework. The EFMM achieved a mean accuracy of 98.38% ± 0.36 and a mean macro-F1 score of 98.48% ± 0.43. Friedman statistical testing demonstrated that there were significant differences between the performance of the classifiers evaluated (p < 0.001), which further validated the performance of the EFMM. Conclusions: To sum up, the proposed SqueezeNet–PCA–EFMM is an effective and efficient method for Alzheimer’s disease stage classification under MRI images. The combination of SqueezeNet, PCA, and EFMM—led not only to high classification performance, but also to good cross validation results. Furthermore, this property of incremental learning is the intrinsic one of the EFMM and renders this framework interesting for its incorporation in the next-generation intelligent clinical decision supports in particular, as medical care evolves. Full article

Quantum cyber-physical systems can involve computation-based strategies like quantum algorithms, physical parts, and communication-based infrastructures such as physical quantum bits (qubits), measurement units, and communication links. A photonics-driven quantum cyber-physical system can rely on physical parameters of the environment or communication links such as the refractive index. The injection of false data into the actual value of the refractive index can cause the use of an incorrect value of the refractive index in the system if needed, and as a result, an incorrect interpretation of information or even access to non-real information instead of the actual information. The first attempt to avoid this issue can be the detection of the existence of the cyber-attacks in the system. This paper will address this challenge using artificial intelligence for binary classification to detect the cyber-attacks in the system. For the proof of concept, the proposed strategy is examined deploying the real and imaginary parts of the refractive index value corresponding to a semiconductor, i.e., GaAs. Different scenarios are performed, including a single evaluation, an analysis of several training runs, a comparison considering two types of normalization techniques, variations in the size of artificial intelligence, and a comparison among different machine learning techniques, including artificial neural networks, decision tree models, a logistic regression model, and support vector machine classifiers. Based on the obtained results, for the single evaluation, the class of 95.24 % of the testing samples could be classified successfully. In addition, for the case of the analysis of several training runs, a total of 9000 runs were run for 60 shallow artificial neural networks with different sizes. For 58 neural networks, the maximum achieved accuracy was 100 %. Besides, for the case of the comparison between the normalization techniques, two methods were evaluated, i.e., min-max and z-score normalization. The results were very close to each other, but, more accurately, z-score normalization indicated a better performance and a higher accuracy. Finally, among the mentioned machine learning models, artificial neural networks mostly showed higher accuracies. Full article

Precise step length estimation (SLE) is a key necessity for not only navigation systems design but also gait health monitoring in neurological conditions. Among existing solutions, non-invasive inertial sensor-based approaches operating without dedicated infrastructure are more cost-effective. Many such methods, however, rely on bodily affixed inertial sensors rather than freely held smartphone sensors. Traditional signal processing approaches, on the other hand, offer varying accuracy across diverse gait patterns due to user parameter calibration. This study, thus, proposes a regression-based SLE framework employing eight regression algorithms: linear regression (LR), k-nearest neighbors, support vector machine, decision tree, elastic network, random forest, histogram-based gradient boosting (HGB) regressor, and artificial neural network (ANN). Their extensive and rigorous evaluations across varied window sizes, using a dataset collected in normal and fast walking modes with two device positions (hand-held and trouser-pocket) during three evaluation scenarios, demonstrate the HGB regressor’s outstanding performance, achieving the lowest mean absolute error (MAE) below 1 cm across four different contexts under leave-one-out cross-validation-based evaluation and three in the seen test evaluations. Moreover, the findings report the ANN’s exceptional generalization capacity over other models and the previous method IRT-SD-SLE in unseen test evaluations, with an MAE not exceeding $6.3$ cm. The extensive evaluations of training and testing times reveal the highest computational efficiency for LR, moderate efficiency for the HGB regressor, and the highest training cost for the ANN, indicating a clear trade-off between MAE and computational expense. Additionally, this study includes an insightful discussion on the performance results, including the trade-offs between accuracy and efficiency. Full article

The achievement of continuous improvement in maintenance, and consequently in its overall efficiency, becomes a cornerstone of railway transport systems, particularly when one of the objectives is to increase operational speeds. Therefore, methodologies that enable the early detection of defects in the most critical components of the system—thus ensuring maximum availability of railway rolling stock while reducing maintenance and operational costs—are of paramount importance. Recently, the feasibility of using decision trees to classify the condition of a bogie wheelset has been analysed through the study of vibration signals obtained from the wheelset on a test bench. These signals were, in turn, decomposed into their intrinsic mode functions, to which classical signal processing techniques were applied, yielding excellent results. The present work therefore constitutes both a continuation and a complement to the aforementioned study, with the ultimate aim of establishing a comparison as well as enhancing the methodology, which in itself represents the principal contribution of this manuscript. Full article

The objective of this study was to predict seed and oil yields in safflower (Carthamus tinctorius L.) using UAV-based multispectral imagery and machine learning algorithms. The study was conducted during the 2024 growing season under varying irrigation levels, fertilization practices, and applications of plant growth-promoting rhizobacteria. The experiment included four irrigation levels, fertilized and unfertilized conditions, and bacterial treatments consisting of Bacillus pumilus, Bacillus albus, their mixture, and a non-bacterial control. Sixty-two vegetation indices were calculated from multispectral images acquired during the harvest maturity period and used to predict seed and oil yields. To identify the most informative features, Mutual Information, Recursive Feature Elimination, and LASSO feature selection methods were applied; subsequently, Linear Regression, Decision Tree, Random Forest, Support Vector Regression, K-Nearest Neighbors, and XGBoost regression algorithms were compared. The results showed that Linear Regression combined with Mutual Information-based feature selection was the most successful approach for predicting both seed and oil yields. According to the 5-fold cross-validation results, average values of $R=0.8600$, $MAE=0.2870$, and $RMSE=0.3765$ were obtained for seed yield prediction, while average values of $R=0.8710$, $MAE=0.0876$, and $RMSE=0.1101$ were obtained for oil yield prediction. Full article

Sentencing is among the most contested tasks in criminal adjudication, as judges must balance retribution, rehabilitation, deterrence, and proportionality within broad judicial discretion. In Taiwan, Article 57 of the Criminal Code mandates consideration of multiple statutory factors, yet substantial inter-case variation persists. This study proposes a framework integrating large language models (LLMs) with traditional machine learning (tree-ensemble regressors) to predict imprisonment duration. Using a question-driven labeling pipeline derived from Article 57, LLMs extract sentencing factors from unstructured judgments and transform them into transparent, auditable feature vectors for regression models. The results show that stronger LLM architectures improve downstream prediction: under the same 72-question and Gradient Boosting Regressor with Optuna setting, Qwen3-30B-A3B reduced MAE by 42.8% and RMSE by 41.6% relative to DeepSeek-R1-Distill-Qwen-32B. Additionally, recursive feature elimination reduced the questionnaire from 72 to 28 items (61.11% reduction) with only a 1.00% increase in MAE and a 1.57% increase in RMSE. The proposed system serves as a judicial decision-support tool providing reference values and consistency checks while preserving due process, transparency, and judicial discretion. Full article

In this work, we investigate the application of Machine Learning (ML) algorithms to the identification and quantitative characterization of quantum entanglement in polarization-entangled photon pairs. The analysis is based on simulated symmetric, informationally complete, positive operator-valued measure (SIC-POVM) measurement data, where each two-qubit state is represented by a 16-dimensional measurement vector corresponding to experimentally accessible coincidence counts. The generated SIC-POVM measurement data include Poissonian shot noise. Several supervised ML algorithms, including Logistic Regression, k-Nearest Neighbors, Decision Trees, Support Vector Machines, and Random Forests, are applied to the classification of separable and entangled states directly from raw measurement data, without explicit density matrix reconstruction or the use of conventional separability criteria. The study additionally explores clustering methods and nonlinear regression techniques for estimating continuous entanglement measures. The obtained results demonstrate that ML methods can achieve very high classification accuracy, even under extremely limited training conditions. These findings indicate that ML may provide an efficient alternative to conventional quantum-state analysis under simulated Poissonian noise conditions. Full article