A comprehensive survey on support vector machine classification: Applications, challenges and trends

Tensor dictionary-based heterogeneous transfer learning to study emotion-related gender differences in brain

In practice, collecting auxiliary labeled data with same feature space from multiple domains is difficult. Thus, we focus on the heterogeneous transfer learning to address the problem of insufficient sample sizes in neuroimaging. Viewing subjects, time, and features as dimensions, brain activation and dynamic functional connectivity data can be treated as high-order heterogeneous data with heterogeneity arising from distinct feature space. To use the heterogeneous priori knowledge from the low-dimensional brain activation data to improve the classification performance of high-dimensional dynamic functional connectivity data, we propose a tensor dictionary-based heterogeneous transfer learning framework. It combines supervised tensor dictionary learning with heterogeneous transfer learning for enhance high-order heterogeneous knowledge sharing. The former can encode the underlying discriminative features in high-order data into dictionaries, while the latter can transfer heterogeneous knowledge encoded in dictionaries through feature transformation derived from mathematical relationship between domains. The primary focus of this paper is gender classification using fMRI data to identify emotion-related brain gender differences during adolescence. Additionally, experiments on simulated data and EEG data are included to demonstrate the generalizability of the proposed method. Experimental results indicate that incorporating prior knowledge significantly enhances classification performance. Further analysis of brain gender differences suggests that temporal variability in brain activity explains differences in emotion regulation strategies between genders. By adopting the heterogeneous knowledge sharing strategy, the proposed framework can capture the multifaceted characteristics of the brain, improve the generalization of the model, and reduce training costs. Understanding the gender specific neural mechanisms of emotional cognition helps to develop the gender-specific treatments for neurological diseases.

Local interpretable model-agnostic explanation approach for medical imaging analysis: A systematic literature review

BACKGROUND

The interpretability and explainability of machine learning (ML) and artificial intelligence systems are critical for generating trust in their outcomes in fields such as medicine and healthcare. Errors generated by these systems, such as inaccurate diagnoses or treatments, can have serious and even life-threatening effects on patients. Explainable Artificial Intelligence (XAI) is emerging as an increasingly significant area of research nowadays, focusing on the black-box aspect of sophisticated and difficult-to-interpret ML algorithms. XAI techniques such as Local Interpretable Model-Agnostic Explanations (LIME) can give explanations for these models, raising confidence in the systems and improving trust in their predictions. Numerous works have been published that respond to medical problems through the use of ML models in conjunction with XAI algorithms to give interpretability and explainability. The primary objective of the study is to evaluate the performance of the newly emerging LIME techniques within healthcare domains that require more attention in the realm of XAI research.

METHOD

A systematic search was conducted in numerous databases (Scopus, Web of Science, IEEE Xplore, ScienceDirect, MDPI, and PubMed) that identified 1614 peer-reviewed articles published between 2019 and 2023.

RESULTS

52 articles were selected for detailed analysis that showed a growing trend in the application of LIME techniques in healthcare, with significant improvements in the interpretability of ML models used for diagnostic and prognostic purposes.

CONCLUSION

The findings suggest that the integration of XAI techniques, particularly LIME, enhances the transparency and trustworthiness of AI systems in healthcare, thereby potentially improving patient outcomes and fostering greater acceptance of AI-driven solutions among medical professionals.

A multi-class support vector machine classification model based on 14 microRNAs for forensic body fluid identification

MicroRNAs (miRNAs) are promising biomarkers for forensic body fluid identification owing to their small size, stability against degradation, and differential expression patterns. However, the expression of most body fluid-miRNAs is relative (differentially expressed in certain body fluids) rather than absolute (exclusively expressed in a specific body fluid). Moreover, different body fluids contain heterogeneous cell types, complicating their identification. Therefore, appropriate normalization strategies to eliminate non-biological variations and robust models to interpret expression levels accurately are necessary prerequisites for applying miRNAs in body fluid identification. In this study, the expression stability of six candidate reference genes (RGs) across five body fluids was validated using geNorm, NormFinder, BestKeeper and RankAggreg, and the most suitable combination of RGs (hsa-miR-484 and hsa-miR-191-5p) was identified under our experimental conditions. Subsequently, we systematically evaluated the expression patterns of the 28 most promising body fluid-specific miRNA markers using TaqMan RT-qPCR and selected the optimal combination of markers (12 miRNAs) to establish a multi-class support vector machine (MSVM) classification model. An independent test set (60 samples) was used to validate the accuracy of the proposed classification model, while an additional 30 casework samples were used to assess its robustness. The MSVM model accurately predicted the body fluid origin for almost all (59/60) single-source samples. Moreover, this model demonstrated the capability to identify aged forensic samples and to predict the primary components of mixed stains to a certain extent. In summary, this study presented a miRNA-based MSVM classification model for forensic body fluid identification using the qPCR platform. However, extensive validation, especially inter-laboratory collaborative exercises, is necessary before miRNA can be routinely applied in forensic identification practice.

Machine learning for classifying chronic ankle instability based on ankle strength, range of motion, postural control and anatomical deformities in delivery service workers with a history of lateral ankle sprains

OBJECTIVE

Chronic ankle instability (CAI) frequently develops as a result of lateral ankle sprains (LAS) in delivery service workers (DSWs). Identifying risk factors for CAI is crucial for implementing targeted interventions. This study aimed to develop machine learning (ML) models for classifying CAI in DSWs with a history of LAS (DSWsLAS) and to identify key contributory factors.

DESIGN

Exploratory, cross-sectional design.

SETTING

and participants: A total of 121 DSWsLAS were screened for eligibility among 289 DSWs.

METHODS

A total of 121 DSWsLAS were assessed for demographic characteristics, including ankle strength, range of motion, postural control, and anatomical deformities. Seven ML algorithms were trained and tested for classifying CAI. Principal component analysis (PCA) was used for feature extraction, and feature permutation importance (FPI) and Shapley additive explanations (SHAP) were employed to identify influential features.

MAIN OUTCOME MEASURES

Model performances were assessed using area under the curve (AUC). To interpret the classifications, we used FPI and SHAP values.

RESULTS

PCA derived 7 principal components (PCs) accounting for 83.5% of the total variation in the data. The support vector machine (SVM) algorithm achieved the highest classifying performance (AUC = 0.817) among the ML models. FPI and SHAP revealed that PC1, PC2, PC5, and PC7 were the most influential features for classifying CAI in DSWsLAS.

CONCLUSIONS

The SVM algorithm, utilizing PCA-derived factors related to body mass index and ankle muscle strength demonstrated high classifying performance for diagnosis of CAI in DSWsLAS, emphasizing the importance of considering multiple contributory factors in the prevention and management of this condition.

Neural Network-Enabled Multiparametric Impedance Signal Templating for High throughput Single-Cell Deformability Cytometry Under Viscoelastic Extensional Flows

Cellular biophysical metrics exhibit systematic alterations during processes, such as metastasis and immune cell activation, which can be used to identify and separate live cell subpopulations for targeting drug screening. Image-based biophysical cytometry under extensional flows can accurately quantify cell deformability based on cell shape alterations but needs extensive image reconstruction, which limits its inline utilization to activate cell sorting. Impedance cytometry can measure these cell shape alterations based on electric field screening, while its frequency response offers functional information on cell viability and interior structure, which are difficult to discern by imaging. Furthermore, 1-D temporal impedance signal trains exhibit characteristic shapes that can be rapidly templated in near real-time to extract single-cell biophysical metrics to activate sorting. We present a multilayer perceptron neural network signal templating approach that utilizes raw impedance signals from cells under extensional flow, alongside its training with image metrics from corresponding cells to derive net electrical anisotropy metrics that quantify cell deformability over wide anisotropy ranges and with minimal errors from cell size distributions. Deformability and electrical physiology metrics are applied in conjunction on the same cell for multiparametric classification of live pancreatic cancer cells versus cancer associated fibroblasts using the support vector machine model.

Smart home-assisted anomaly detection system for older adults: a deep learning approach with a comprehensive set of daily activities

Smart homes have the potential to enable remote monitoring of the health and well-being of older adults, leading to improved health outcomes and increased independence. However, current approaches only consider a limited set of daily activities and do not combine data from individuals. In this work, we propose the use of deep learning techniques to model behavior at the population level and detect significant deviations (i.e., anomalies) while taking into account the whole set of daily activities (41). We detect and visualize daily routine patterns, train a set of recurrent neural networks for behavior modelling with next-day prediction, and model errors with a normal distribution to identify significant deviations while considering the temporal component. Clustering of daily routines achieves a silhouette score of 0.18 and the best model obtains a mean squared error in next day routine prediction of 4.38%. The mean number of deviated activities for the anomalies in the train and test set are 3.6 and 3.0, respectively, with more than 60% of anomalies involving three or more deviated activities in the test set. The methodology is scalable and can incorporate additional activities into the analysis.

Temperature-Enhanced Purine Metabolism-Based Versatile SERS Platform for Rapid Clinical Pathogens Diagnosis and Drug-Resistant Assessment

Label-free surface-enhanced Raman spectroscopy (SERS) combined with machine learning (ML) techniques presents a promising approach for rapid pathogen identification. Previous studies have demonstrated that purine degradation metabolites are the primary contributors to SERS spectra; however, generating these distinguishable spectra typically requires a long incubation time (>10 h) at room temperature. Moreover, the lack of attention to spectral variations between strains of the same bacterial species has limited the generalizability of ML models in real-world applications. To address these issues, we investigated temperature-induced alterations in bacterial purine metabolism and found that robust SERS spectra could be obtained within just 1 h by heating samples to 60 °C. Our study further revealed that pathogens exhibit multiple fingerprint patterns across strains, rather than a uniform spectral signature. To enhance practicality, we optimized ML models by training them on data sets capturing all relevant SERS fingerprints and validated them on separate bacterial strains. The SoftMax classifier achieved 100% accuracy in identifying both laboratory and clinical specimens within 17 h. Additionally, the platform demonstrated over 91% accuracy in distinguishing drug-resistant strains, such as methicillin-resistant Staphylococcus aureus and carbapenem-resistant Klebsiella pneumoniae, and achieved 99.66% accuracy in differentiating specific strains within a species, such as enterohemorrhagic Escherichia coli. This accelerated, purine metabolism-based SERS platform offers a highly promising alternative for the rapid diagnosis of bacterial infections.

Nonlinear Small Sample Data Regression with a New Rational-Quadratic Minkowski Kernel for Tobacco Laser Perforation Process Tar Reduction Estimation

This paper investigates the nonlinear relationship between tobacco harmful content tar reduction and laser perforation parameters. To find a model to demonstrate the relationship between the laser perforation parameters and the cigarette tar reduction level, an online platform based on Python Streamlit was built to collect and publish related data. After the initial analysis of the collected experimental data, the quadratic nonlinear regression model demonstrates a significant fit to the experimental data. However, although the nonlinear regression has much higher accuracy than the linear regression plane, the prediction normalized root mean squared error (NRMSE) is still high, over 10%, which indicates that the regression relationship is more complex than the simple quadratic function expression. On the other hand, the sample dataset used for modeling is very limited, which restricts its exploration and the development of a model comparable to those built with big data. To address this challenge for small sample size data in modeling this complex nonlinear relationship, a novel rational-quadratic Minkowski (RM)-based kernel was designed. This RM-kernel model acquires higher accuracy than other kernels in both SVM and Gaussian process regression. Furthermore, this new kernel also shows less sensitivity to hyperparameter change, the greater ability to capture complex relationships, and more flexibility than the RBF kernel and RQ kernel. Subsequently, the kernel-based RM regression model was successfully implemented for laser perforation parameter selection, yielding consistent results that align with human sensory test data.

Minimal sourced and lightweight federated transfer learning models for skin cancer detection

One of the most fatal diseases that affect people is skin cancer. Because nevus and melanoma lesions are so similar and there is a high likelihood of false negative diagnoses challenges in hospitals. The aim of this paper is to propose and develop a technique to classify type of skin cancer with high accuracy using minimal resources and lightweight federated transfer learning models. Here minimal resource based pre-trained deep learning models including EfficientNetV2S, EfficientNetB3, ResNet50, and NasNetMobile have been used to apply transfer learning on data of shape[Formula: see text]. To compare with applied minimal resource transfer learning, same methodology has been applied using best identified model i.e. EfficientNetV2S for images of shape[Formula: see text]. The identified minimal and lightweight resource based EfficientNetV2S with images of shape [Formula: see text] have been applied for federated learning ecosystem. Both, identically and non-identically distributed datasets of shape [Formula: see text] have been applied and analyzed through federated learning implementations. The results have been analyzed to show the impact of low-pixel images with non-identical distributions over clients using parameters such as accuracy, precision, recall and categorical losses. The classification of skin cancer shows an accuracy of IID 89.83% and Non-IID 90.64%.

Improved BCI calibration in multimodal emotion recognition using heterogeneous adversarial transfer learning

The use of brain-computer interface (BCI) technology to identify emotional states has gained significant interest, especially with the rise of virtual reality (VR) applications. However, the extensive calibration required for precise emotion recognition models presents a significant challenge, particularly for sensitive groups such as children, elderly, and patients. This study presents a novel approach that utilizes heterogeneous adversarial transfer learning (HATL) to synthesize electroencephalography (EEG) data from various other signal modalities, reducing the need for lengthy calibration phases. We benchmark the efficacy of three generative adversarial network (GAN) architectures, such as conditional GAN (CGAN), conditional Wasserstein GAN (CWGAN), and CWGAN with gradient penalty (CWGAN-GP) within this framework. The proposed framework is rigorously tested on two conventional open sourced datasets, SEED-V and DEAP. Additionally, the framework was applied to an immersive three-dimensional (3D) dataset named GraffitiVR, which we collected to capture the emotional and behavioral reactions of individuals experiencing urban graffiti in a VR environment. This expanded application provides insights into emotion recognition frameworks in VR settings, providing a wider range of contexts for assessing our methodology. When the accuracy of emotion recognition classifiers trained with CWGAN-GP-generated EEG data combined with non-EEG sensory data was compared against those trained using a combination of real EEG and non-EEG sensory data, the accuracy ratios were 93% on the SEED-V dataset, 99% on the DEAP dataset, and 97% on the GraffitiVR dataset. Moreover, in the GraffitiVR dataset, using CWGAN-GP-generated EEG data with non-EEG sensory data for emotion recognition models resulted in up to a 30% reduction in calibration time compared to classifiers trained on real EEG data with non-EEG sensory data. These results underscore the robustness and versatility of the proposed approach, significantly enhancing emotion recognition processes across a variety of environmental settings.

Artificial intelligence for the detection of acute myeloid leukemia from microscopic blood images; a systematic review and meta-analysis

Background

Leukemia is the 11th most prevalent type of cancer worldwide, with acute myeloid leukemia (AML) being the most frequent malignant blood malignancy in adults. Microscopic blood tests are the most common methods for identifying leukemia subtypes. An automated optical image-processing system using artificial intelligence (AI) has recently been applied to facilitate clinical decision-making.

Aim

To evaluate the performance of all AI-based approaches for the detection and diagnosis of acute myeloid leukemia (AML).

Methods

Medical databases including PubMed, Web of Science, and Scopus were searched until December 2023. We used the "metafor" and "metagen" libraries in R to analyze the different models used in the studies. Accuracy and sensitivity were the primary outcome measures.

Results

Ten studies were included in our review and meta-analysis, conducted between 2016 and 2023. Most deep-learning models have been utilized, including convolutional neural networks (CNNs). The common- and random-effects models had accuracies of 1.0000 [0.9999; 1.0001] and 0.9557 [0.9312, and 0.9802], respectively. The common and random effects models had high sensitivity values of 1.0000 and 0.8581, respectively, indicating that the machine learning models in this study can accurately detect true-positive leukemia cases. Studies have shown substantial variations in accuracy and sensitivity, as shown by the Q values and I2 statistics.

Conclusion

Our systematic review and meta-analysis found an overall high accuracy and sensitivity of AI models in correctly identifying true-positive AML cases. Future research should focus on unifying reporting methods and performance assessment metrics of AI-based diagnostics.

Systematic review registration

https://www.crd.york.ac.uk/prospero/#recordDetails, CRD42024501980.

Alzheimer's disease diagnosis using rhythmic power changes and phase differences: a low-density EEG study

Objectives

The future emergence of disease-modifying treatments for dementia highlights the urgent need to identify reliable and easily accessible tools for diagnosing Alzheimer's disease (AD). Electroencephalography (EEG) is a non-invasive and cost-effective technique commonly used in the study of neurodegenerative disorders. However, the specific alterations in EEG biomarkers associated with AD remain unclear when using a limited number of electrodes.

Methods

We studied pathological characteristics of AD using low-density EEG data collected from 26 AD and 29 healthy controls (HC) during both eye closed (EC) and eye opened (EO) resting conditions. The analysis including power spectrum, phase lock value (PLV), and weighted lag phase index (wPLI) and power-to-power frequency coupling (theta/beta) analysis were applied to extract features in the delta, theta, alpha, and beta bands.

Results

During the EC condition, the AD group exhibited decreased alpha power compared to HC. Additionally, both analysis of PLV and wPLI in the theta band indicated that the alterations in the AD brain network predominantly involved in the frontal region with the opposite changes. Moreover, the AD group had increased frequency coupling in the frontal and central regions. Surprisingly, no group difference was found in the EO condition. Notably, decreased theta band functional connectivity within the fronto-central lobe and increased frequency coupling in frontal region were found in AD group from EC to EO. More importantly, the combination of EC and EO quantitative EEG features improved the inter-group classification accuracy when using support vector machine (SVM) in older adults with AD. These findings highlight the complementary nature of EC and EO conditions in assessing and differentiating AD cohorts.

Conclusion

Our results underscore the potential of utilizing low-density EEG data from resting-state paradigms, combined with machine learning techniques, to improve the identification and classification of AD.

Protocol for identifying Mycobacterium tuberculosis infection status through airway microbiome profiling

This protocol describes the steps to determine an airway microbiome signature for identifying Mycobacterium tuberculosis infection status. We outline procedures for processing microbiome data, calculating diversity measures, and fitting Dirichlet multinomial mixture models. Additionally, we provide steps for analyzing taxonomic relative and differential abundances, as well as identifying potential biomarkers associated with infection status. For complete details on the use and execution of this protocol, please refer to Kayongo et al.1.

MSBKA: A Multi-Strategy Improved Black-Winged Kite Algorithm for Feature Selection of Natural Disaster Tweets Classification

With the advancement of the Internet, social media platforms have gradually become powerful in spreading crisis-related content. Identifying informative tweets associated with natural disasters is beneficial for the rescue operation. When faced with massive text data, choosing the pivotal features, reducing the calculation expense, and increasing the model classification performance is a significant challenge. Therefore, this study proposes a multi-strategy improved black-winged kite algorithm (MSBKA) for feature selection of natural disaster tweets classification based on the wrapper method's principle. Firstly, BKA is improved by utilizing the enhanced Circle mapping, integrating the hierarchical reverse learning, and introducing the Nelder-Mead method. Then, MSBKA is combined with the excellent classifier SVM (RBF kernel function) to construct a hybrid model. Finally, the MSBKA-SVM model performs feature selection and tweet classification tasks. The empirical analysis of the data from four natural disasters shows that the proposed model has achieved an accuracy of 0.8822. Compared with GA, PSO, SSA, and BKA, the accuracy is increased by 4.34%, 2.13%, 2.94%, and 6.35%, respectively. This research proves that the MSBKA-SVM model can play a supporting role in reducing disaster risk.

A novel non-invasive EEG-SSVEP diagnostic tool for color vision deficiency in individuals with locked-in syndrome

Introduction

Color vision deficiency (CVD), a common visual impairment, affects individuals' ability to differentiate between various colors due to malfunctioning or absent color photoreceptors in the retina. Currently available diagnostic tests require a behavioral response, rendering them unsuitable for individuals with limited physical and communication abilities, such as those with locked-in syndrome. This study introduces a novel, non-invasive method that employs brain signals, specifically Steady-State Visually Evoked Potentials (SSVEPs), along with Ishihara plates to diagnose CVD. This method aims to provide an alternative diagnostic tool that addresses the limitations of current tests.

Methods

Electroencephalography (EEG) recordings were obtained from 16 subjects, including 5 with CVD (specifically Deuteranomaly), using channels O1, O2, Pz, and Cz. The subjects were exposed to visual stimuli at frequencies of 15 Hz and 18 Hz to assess the proposed method. The subjects focused on specific visual stimuli in response to questions related to the Ishihara plates. Their responses were analyzed to determine the presence of CVD. Feature extraction was performed using Power Spectral Density (PSD), Canonical Correlation Analysis (CCA), and a combined PSD + CCA, followed by classification to categorize subjects into two classes: normal vision and CVD.

Results

The results indicate that the proposed method effectively diagnoses CVD in individuals with limited communication abilities. The classification accuracy of SSVEP exceeded 75% across the three classifiers: Decision Tree (DT), K-Nearest Neighbors (KNN), and Support Vector Machine (SVM). The SVM classifier demonstrated higher accuracy compared to the other classifiers, exceeding 90%.

Discussion

These observations suggest that the SVM classifier, utilizing the combined feature set of PSD + CCA, may be the most effective in this classification task. These findings demonstrate that the proposed method is an accurate and reliable diagnostic tool for CVD, particularly for individuals unable to communicate.

Diagnosis of Alzheimer's disease using FusionNet with improved secretary bird optimization algorithm for optimal MK-SVM based on imaging genetic data

Alzheimer's disease is an irreversible central neurodegenerative disease, and early diagnosis of Alzheimer's disease is beneficial for its prevention and early intervention treatment. In this study, we propose a novel framework, FusionNet-ISBOA-MK-SVM, which integrates a fusion network (FusionNet) and improved secretary bird optimization algorithm to optimize multikernel support vector machine for Alzheimer's disease diagnosis. The model leverages multimodality data, including functional magnetic resonance imaging and genetic information (single-nucleotide polymorphisms). Specifically, FusionNet employs U-shaped hierarchical graph convolutional networks and sparse graph attention networks to select feature effectively. Extensive validation using the Alzheimer's Disease Neuroimaging Initiative dataset demonstrates the model's superior interpretability and classification performance. Compared to other state-of-the-art machine learning methods, FusionNet-ISBOA-MK-SVM achieves classification accuracies of 98.6%, 95.7%, 93.0%, 91.8%, 93.1%, and 95.4% for HC vs. AD, EMCI vs. AD, LMCI vs. AD, EMCI vs. AD, HC vs. EMCI, and HC vs. LMCI, respectively. Moreover, the proposed model identifies affected brain regions and pathogenic genes, offering deeper insights into the mechanisms and progression of Alzheimer's disease. These findings provide valuable scientific evidence to support early diagnosis and preventive strategies for Alzheimer's disease.

Support matrix machine: A review

Support vector machine (SVM) is one of the most studied paradigms in the realm of machine learning for classification and regression problems. It relies on vectorized input data. However, a significant portion of the real-world data exists in matrix format, which is given as input to SVM by reshaping the matrices into vectors. The process of reshaping disrupts the spatial correlations inherent in the matrix data. Also, converting matrices into vectors results in input data with a high dimensionality, which introduces significant computational complexity. To overcome these issues in classifying matrix input data, support matrix machine (SMM) is proposed. It represents one of the emerging methodologies tailored for handling matrix input data. SMM preserves the structural information of the matrix data by using the spectral elastic net property which is a combination of the nuclear norm and Frobenius norm. This article provides the first in-depth analysis of the development of the SMM model, which can be used as a thorough summary by both novices and experts. We discuss numerous SMM variants, such as robust, sparse, class-imbalance, and multi-class classification models. We also analyze the applications of the SMM and conclude the article by outlining potential future research avenues and possibilities that may motivate researchers to advance the SMM algorithm.

Progress of machine learning-based biosensors for the monitoring of food safety: A review

Rapid urbanization and growing food demand caused people to be concerned about food safety. Biosensors have gained considerable attention for assessing food safety due to selectivity, and sensitivity but poor stability inherently limits their application. The emergence of machine learning (ML) has enhanced the efficiency of different sensors for food safety assessment. The ML combined with various noninvasive biosensors has been implemented efficiently to monitor food safety by considering the stability of bio-recognition molecules. This review comprehensively summarizes the application of ML-powered biosensors to investigate food safety. Initially, different detector-based biosensors using biological molecules with their advantages and disadvantages and biosensor-related various ML algorithms for food safety monitoring have been discussed. Next, the application of ML-powered biosensors to detect antibiotics, foodborne microorganisms, mycotoxins, pesticides, heavy metals, anions, and persistent organic pollutants has been highlighted for the last five years. The challenges and prospects have also been deliberated. This review provides a new prospect in developing various biosensors for multi-food contaminants powered by suitable ML algorithms to monitor in-situ food safety.

Laser Desorption-Ion Mobility Spectrometry of Explosives for Forensic and Security Applications

BACKGROUND

The detection of explosives in crime scene investigations is critical for forensic science. This study explores the application of laser desorption (LD) ion mobility spectrometry (IMS) as a novel method for this purpose utilising a new IMS prototype developed by MaSaTECH.

METHODS

The LD sampling technique employs a laser diode module to vaporise explosive traces on surfaces, allowing immediate analysis by IMS without sample preparation. Chemometric approaches, including multivariate data analysis, were utilised for data processing and interpretation, including pre-processing of raw IMS plasmagrams and various pattern recognition techniques, such as linear discriminant analysis (LDA) and support vector machines (SVMs).

RESULTS

The IMS prototype was validated through experiments with pure explosives (TNT, RDX, PETN) and explosive products (SEMTEX 1A, C4) on different materials. The study found that the pre-processing method significantly impacts classification accuracy, with the PCA-LDA model demonstrating the best performance for real-world applications.

CONCLUSIONS

The LD-IMS prototype, coupled with effective chemometric techniques, presents a promising methodology for the detection of explosives in forensic investigations, enhancing the reliability of field applications.

An Instrumented Hammer to Detect the Rupture of the Pterygoid Plates

PURPOSE

Craniofacial osteotomies involving pterygomaxillary disjunction are common procedures in maxillofacial surgery. Surgeons still rely on their proprioception to determine when to stop impacting the osteotome, which is important to avoid complications such as dental damage and bleeding. Our group has developed a technique consisting in using an instrumented hammer that can provide information on the mechanical properties of the tissue located around the osteotome tip. The aim of this study is to determine whether a mallet instrumented with a force sensor can be used to predict the crossing of the osteotome through the pterygoid plates.

METHODS

31 osteotomies were carried out in 16 lamb skulls. For each impact, the force signal obtained was analysed using a dedicated signal processing technique. A prediction algorithm based on an SVM classifier and a cost matrix was applied to the database.

RESULTS

We showed that the device could always detect the crossing of the osteotome, sometimes before its occurrence. The prediction accuracy of the device was 94.7%. The method seemed to be sensitive to the thickness of the plate and to crack apparition and propagation.

CONCLUSION

These results pave the way for the development of a per-operative decision support system in maxillofacial surgery.

α-decay half-life predictions with support vector machine

In this study, we investigate the application of support vector machines utilizing a radial basis function kernel for predicting nuclear α-decay half-lives. Our approach integrates a comprehensive set of physics-derived features, including characteristics derived from nuclear structure, to systematically evaluate their impact on predictive accuracy. In addition to traditional parameters such as proton and neutron numbers, as well as terms based on the liquid drop model (e.g., volume, surface, Coulomb features), we incorporate decay energies and orbital angular momentum quantum numbers for both parent and daughter nuclei. Our analysis of 2232 nuclear data points demonstrates that the use of the radial basis function kernel yields predictive models with root mean square errors of 0.819 (for set1) and 0.352 (for set2), aligning with results obtained from comparable machine learning methodologies. Furthermore, Shapley additive explanations values highlight the predominant role of parent nuclei in predicting α-decay half-lives within the support vector machines. These results highlight the effectiveness of machine learning in nuclear structure research, opening up new possibilities for predicting the α-decay half-lives of previously unstudied nuclei.

A Generalized Multi-Detector Combination Approach for Differential Item Functioning Detection

Many studies on differential item functioning (DIF) detection rely on single detection methods (SDMs), each of which necessitates specific assumptions that may not always be validated. Using an inappropriate SDM can lead to diminished accuracy in DIF detection. To address this limitation, a novel multi-detector combination (MDC) approach is proposed. Unlike SDMs, MDC effectively evaluates the relevance of different SDMs under various test conditions and integrates them using supervised learning, thereby mitigating the risk associated with selecting a suboptimal SDM for DIF detection. This study aimed to validate the accuracy of the MDC approach by applying five types of SDMs and four distinct supervised learning methods in MDC modeling. Model performance was assessed using the area under the curve (AUC), which provided a comprehensive measure of the ability of the model to distinguish between classes across all threshold levels, with higher AUC values indicating higher accuracy. The MDC methods consistently achieved higher average AUC values compared to SDMs in both matched test sets (where test conditions align with the training set) and unmatched test sets. Furthermore, MDC outperformed all SDMs under each test condition. These findings indicated that MDC is highly accurate and robust across diverse test conditions, establishing it as a viable method for practical DIF detection.

Application of machine learning for mass spectrometry-based multi-omics in thyroid diseases

Thyroid diseases, including functional and neoplastic diseases, bring a huge burden to people's health. Therefore, a timely and accurate diagnosis is necessary. Mass spectrometry (MS) based multi-omics has become an effective strategy to reveal the complex biological mechanisms of thyroid diseases. The exponential growth of biomedical data has promoted the applications of machine learning (ML) techniques to address new challenges in biology and clinical research. In this review, we presented the detailed review of applications of ML for MS-based multi-omics in thyroid disease. It is primarily divided into two sections. In the first section, MS-based multi-omics, primarily proteomics and metabolomics, and their applications in clinical diseases are briefly discussed. In the second section, several commonly used unsupervised learning and supervised algorithms, such as principal component analysis, hierarchical clustering, random forest, and support vector machines are addressed, and the integration of ML techniques with MS-based multi-omics data and its application in thyroid disease diagnosis is explored.

Machine Learning for Enhanced COPD Diagnosis: A Comparative Analysis of Classification Algorithms

Background: In the United States, chronic obstructive pulmonary disease (COPD) is a significant cause of mortality. As far as we know, it is a chronic, inflammatory lung condition that cuts off airflow to the lungs. Many symptoms have been reported for such a disease: breathing problems, coughing, wheezing, and mucus production. Patients with COPD might be at risk, since they are more susceptible to heart disease and lung cancer. Methods: This study reviews COPD diagnosis utilizing various machine learning (ML) classifiers, such as Logistic Regression (LR), Gradient Boosting Classifier (GBC), Support Vector Machine (SVM), Gaussian Naïve Bayes (GNB), Random Forest Classifier (RFC), K-Nearest Neighbors Classifier (KNC), Decision Tree (DT), and Artificial Neural Network (ANN). These models were applied to a dataset comprising 1603 patients after being referred for a pulmonary function test. Results: The RFC has achieved superior accuracy, reaching up to 82.06% in training and 70.47% in testing. Furthermore, it achieved a maximum F score in training and testing with an ROC value of 0.0.82. Conclusions: The results obtained with the utilized ML models align with previous work in the field, with accuracies ranging from 67.81% to 82.06% in training and from 66.73% to 71.46% in testing.

Effect of Daily Vitamin C Supplementation with or Without Flavonoids on Periodontal, Microbial, and Systemic Conditions Before and After Periodontal Therapy: A Case Series from an RCT

Purpose: To investigate the effect of vitamin C supplementation with or without flavonoids on periodontal conditions, and microbial and systemic variables before and after non-surgical periodontal treatment (NSPT). Materials and Methods: A case series derived from a randomized controlled trial was conducted to explore the effects of daily vitamin C supplementation, with or without flavonoids, on periodontal conditions. The study population was recruited from patients with periodontitis who had been referred to the Department of Periodontology at the Academic Centre for Dentistry Amsterdam (ACTA). The study consisted of a 2-month observation of untreated periodontitis followed by a 3-month period after NSPT. Descriptive statistics, correlation and clustering analyses, and dimensionality reduction methods were used to evaluate the interventions' impact. Results: Due to COVID-19, the study was prematurely terminated and reported findings from 13 patients. Results indicate a correlation between higher plasma vitamin C levels and reduced gingival inflammation, suggesting benefits for untreated periodontal conditions. Clustering analysis showed no differences based on supplementation type, indicating it did not affect outcomes, and microbiological data had limited effects. Principal Component Analysis visualized clusters and illustrated no distinct groups corresponding to supplementation types. Violin plots highlighted variability, with one cluster comprising individuals with more severe periodontal conditions. Conclusions: Higher plasma vitamin C levels were associated with lower gingival inflammation. However, daily vitamin C supplementation, with or without flavonoids, did not show additional benefits on periodontal conditions before or after treatment. Clustering suggests that periodontal severity, rather than supplementation, influenced patient profiles. The study's small sample size limits the generalizability of the findings.

ATR-FTIR spectroscopy and machine/deep learning models for detecting adulteration in coconut water with sugars, sugar alcohols, and artificial sweeteners

Packaged coconut water offers various options, from pure to those with added sugars and other additives. While the purity of coconut water is esteemed for its health benefits, its popularity also exposes it to potential adulteration and misrepresentation. To address this concern, our study combines Fourier transform infrared spectroscopy (FTIR) and machine learning techniques to detect potential adulterants in coconut water through classification models. The dataset comprises infrared spectra from coconut water samples spiked with 15 different types of potential sugar substitutes, including: sugars, artificial sweeteners, and sugar alcohols. The interaction of infrared light with molecular bonds generates unique molecular fingerprints, forming the basis of our analysis. Departing from previous research predominantly reliant on linear-based chemometrics for adulterant detection, our study explored linear, non-linear, and combined feature extraction models. By developing an interactive application utilizing principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE), non-targeted sugar adulterant detection was streamlined through enhanced visualization and pattern recognition. Targeted analysis using ensemble learning random forest (RF) and deep learning 1-dimensional convolutional neural network (1D CNN) achieved higher classification accuracies (95% and 96%, respectively) compared to sparse partial least squares discriminant analysis (sPLS-DA) at 77% and support vector machine (SVM) at 88% on the same dataset. The CNN's demonstrated classification accuracy is complemented by exceptional efficiency through its ability to train and test on raw data.

Artificial intelligence in antidiabetic drug discovery: The advances in QSAR and the prediction of α-glucosidase inhibitors

Artificial Intelligence is transforming drug discovery, particularly in the hit identification phase of therapeutic compounds. One tool that has been instrumental in this transformation is Quantitative Structure-Activity Relationship (QSAR) analysis. This computer-aided drug design tool uses machine learning to predict the biological activity of new compounds based on the numerical representation of chemical structures against various biological targets. With diabetes mellitus becoming a significant health challenge in recent times, there is intense research interest in modulating antidiabetic drug targets. α-Glucosidase is an antidiabetic target that has gained attention due to its ability to suppress postprandial hyperglycaemia, a key contributor to diabetic complications. This review explored a detailed approach to developing QSAR models, focusing on strategies for generating input variables (molecular descriptors) and computational approaches ranging from classical machine learning algorithms to modern deep learning algorithms. We also highlighted studies that have used these approaches to develop predictive models for α-glucosidase inhibitors to modulate this critical antidiabetic drug target.

Negative Emotion Differentiation Promotes Cognitive Reappraisal: Evidence From Electroencephalogram Oscillations and Phase-Amplitude Coupling

Cognitive reappraisal, an effective emotion regulation strategy, is influenced by various individual factors. Although previous studies have established a link between negative emotion differentiation (NED) and cognitive reappraisal, the underlying neural mechanisms remain largely unknown. Using electroencephalography, this study investigates the influence and neural basis of NED in cognitive reappraisal by integrating aspects of event-related potentials, neural oscillation rhythms, and cross-frequency coupling. The findings revealed that individuals with high NED demonstrated a significant decrease in parietal late positive potential amplitudes during cognitive reappraisal, suggesting enhanced cognitive reappraisal abilities. Moreover, high NED individuals displayed increased γ synchronization, parietal α-γ coupling, and frontal θ-γ coupling when reappraising negative emotions than those with low emotion differentiation ability. Machine learning analysis of these neural indicators highlighted the superior classification and predictive accuracy of multimodal indicators for NED as opposed to unimodal indicators. Overall, this multimodal evidence provides a comprehensive interpretation of the neurophysiological mechanisms through which NED influences cognitive reappraisal and provides preliminary empirical support for personalized cognitive reappraisal interventions to alleviate emotional problems.

Machine learning approach for high-throughput phenolic antioxidant screening in black Rice germplasm collection based on surface FTIR

Pigmented rice contains beneficial phenolic antioxidants but analysing them across germplasm collections is laborious and time-consuming. Here we utilised rapid surface Fourier transform infrared (FTIR) spectroscopy and machine learning algorithms (ML) to predict and classify polyphenolic antioxidants. Total phenolics, flavonoids, anthocyanins, and proanthocyanidins were quantified biochemically from 270 diverse global coloured rice collection and attenuated total reflectance (ATR) FTIR spectra were obtained by scanning whole grain surfaces at 800-4000 cm-1. Five ML classification models were optimised using the biochemical and spectral data which performed predictions with 93.5%-100% accuracy. Random Forest and Support Vector Machine models identified key FTIR peaks linked to flavonols, flavones and anthocyanins as important model predictors. This research successfully established direct and non-destructive surface chemistry spectroscopy of the aleurone layer of pigmented rice integrated with ML models as a viable high-throughput platform to accelerate the analysis and profiling of nutritionally valuable coloured rice varieties.

Detection of Escherichia coli Using Bacteriophage T7 and Analysis of Excitation‑Emission Matrix Fluorescence Spectroscopy

Conventional detection methods require the isolation and enrichment of bacteria, followed by molecular, biochemical, or culture-based analysis. To address some of the limitations of conventional methods, this study develops a machine learning (ML) approach to analyze the excitation-emission matrix (EEM) fluorescence data generated based on bacteriophage T7 and Escherichia coli interactions for in-situ detection of live bacteria in the presence of fresh produce homogenate. We trained classification models using various ML algorithms based on the 3-D EEM data generated with bacteria and their interactions with a T7 phage. These ML algorithms, including linear Support Vector Classifier (SVC) and Random Forest (RF), demonstrate high accuracy (>0.85) for detecting E. coli at 102 CFU/ml concentration within 6 h. Additionally, these ML models can differentiate among different E. coli concentration levels. For example, the Gaussian Process model achieved an accuracy of 92% in detecting different concentration levels of live E. coli. Application of these ML methods to detect E. coli in spinach homogenate yielded an accuracy of 89% using the linear-SVC model. Furthermore, feature selection techniques were employed to reduce the dimensionality of the data, revealing that only six features were necessary for achieving classification accuracy (>0.85) of spinach homogenate samples containing 102 CFU/ml of E. coli. These findings highlight the potential of this novel bacterial detection methodology, offering rapid, specific, and efficient solutions for applications in food safety and environmental monitoring.

Prediction of mortality in hemodialysis patients based on autoencoders

BACKGROUND

Patients with end-stage renal disease (ESRD) undergoing hemodialysis (HD) exhibit a high mortality risk, particularly at the onset of treatment. Conventional risk assessment models, dependent on extensive temporal data accumulation, frequently encounter issues of data incompleteness and lengthy collection periods.

OBJECTIVE

This study addresses the imbalance in short-term HD data and the issue of missing data features, achieving a robust assessment of mortality risk for HD patients over the subsequent 30 to 450 days.

METHODS

An autoencoder-based mortality prediction model for HD patients is proposed. Leveraging the manifold structure of the non-missing features and the intrinsic relationship between the features in the high-dimensional data space, the model infers the values of the missing features. Noise and redundant information typically distort the manifold structure, impacting the accuracy of inferences about missing features. Consequently, we generate feature dropping masks to simulate the missing data distribution in the deep learning framework and design an autoencoder, forming an adaptive feature extraction module. The module utilizes readily available short-term data for unsupervised learning, enabling the encoder to reconstruct missing features and derive latent representations. Finally, a classifier based on the latent representations achieves the mortality prediction.

RESULTS

Over a 30-day observation window, the model demonstrated superior mortality prediction performance compared to other models across all prediction windows. Feature importance analysis showed that creatinine and age are consistently the most critical features across all prediction windows. Glucose (fasting) and platelet count also remain significant, with their correlation with mortality risk increasing over time. Serum albumin, international standard ratio, and phosphate are particularly important for short-term predictions, while conjugated bilirubin and prothrombin time gain prominence in mid- and long-term predictions.

CONCLUSION

The proposed model proficiently leverages short-term HD data to provide precise mortality risk evaluations in HD patients, with particular efficacy in the short-term. Its application holds considerable value for clinical decision-making and risk management in this patient population.

Classifying Dry Eye Disease Patients from Healthy Controls Using Machine Learning and Metabolomics Data

Background: Dry eye disease is a common disorder of the ocular surface, leading patients to seek eye care. Clinical signs and symptoms are currently used to diagnose dry eye disease. Metabolomics, a method for analyzing biological systems, has been found helpful in identifying distinct metabolites in patients and in detecting metabolic profiles that may indicate dry eye disease at early stages. In this study, we explored the use of machine learning and metabolomics data to identify cataract patients who suffer from dry eye disease, a topic that, to our knowledge, has not been previously explored. As there is no one-size-fits-all machine learning model for metabolomics data, choosing the most suitable model can significantly affect the quality of predictions and subsequent metabolomics analyses. Methods: To address this challenge, we conducted a comparative analysis of eight machine learning models on two metabolomics data sets from cataract patients with and without dry eye disease. The models were evaluated and optimized using nested k-fold cross-validation. To assess the performance of these models, we selected a set of suitable evaluation metrics tailored to the data set's challenges. Results: The logistic regression model overall performed the best, achieving the highest area under the curve score of 0.8378, balanced accuracy of 0.735, Matthew's correlation coefficient of 0.5147, an F1-score of 0.8513, and a specificity of 0.5667. Additionally, following the logistic regression, the XGBoost and Random Forest models also demonstrated good performance. Conclusions: The results show that the logistic regression model with L2 regularization can outperform more complex models on an imbalanced data set with a small sample size and a high number of features, while also avoiding overfitting and delivering consistent performance across cross-validation folds. Additionally, the results demonstrate that it is possible to identify dry eye in cataract patients from tear film metabolomics data using machine learning models.

Assessment of Tail-Cutting in Frozen Albacore (Thunnus alalunga) Through Ultrasound Inspection and Chemical Analysis

Fat content is the main criterion for evaluating albacore quality. However, no reports exist on the accuracy of the tail-cutting method, a method used to assess the fat content of albacore. Here, we evaluated this method by comparing it with chemical analysis and ultrasound inspection. We measured the actual fat content in albacore using chemical analysis and compared the results with those obtained using the tail-cutting method. Significant discrepancies (99% CI, t-test) were observed in fat content among the tail-cutting samples. Using chemical analysis as the ground truth, the accuracy of tail-cutting from two different companies was 70.0% for company A and 51.9% for company B. An ultrasound inspection revealed that a higher fat content reduced the amplitude of ultrasound signals with statistical significance (99% CI, t-test). Finally, machine learning algorithms were used to enforce the ultrasound inspection. The best combination of ultrasound inspection and a machine learning algorithm achieved an 84.2% accuracy for selecting fat-rich albacore, which is better than tail-cutting (73.6%). Our findings suggested that ultrasound inspection could be a valuable and non-destructive method for estimating the fat content of albacore, achieving better accuracy than the traditional tail-cutting method.

Mapping surface soil organic carbon density of cultivated land using machine learning in Zhengzhou

Research on soil organic carbon (SOC) is crucial for improving soil carbon sinks and achieving the "double-carbon" goal. This study introduces ten auxiliary variables based on the data from a 2021 land quality survey in Zhengzhou and a multi-objective regional geochemical survey. It uses geostatistical ordinary kriging (OK) interpolation, as well as classical machine learning (ML) models, including random forest (RF) and support vector machine (SVM), to map soil organic carbon density (SOCD) in the topsoil layer (0 - 20 cm) of cultivated land. It partitions the sampling data to assess the generalization capability of the machine learning models, with Zhongmu County designated as an independent test set (dataset2) and the remaining data as the training set (dataset1). The three models are trained using dataset1, and the trained machine learning models are directly applied to dataset2 to evaluate and compare their generalization performance. The distribution of SOCD and SOCS in soils of various types and textures is analyzed using the optimal interpolation method. The results indicated that: (1) The average SOC densities predicted by OK interpolation, RF, and SVM are 3.70, 3.74, and 3.63 kg/m2, with test set precisions (R2) of 0.34, 0.60, and 0.81, respectively. (2) ML achieves a significantly higher predictive precision than traditional OK interpolation. The RF model's precision is 0.21 higher than the SVM model and more precise in estimating carbon stock. (3) When applied to the dataset2, the RF model exhibited superior generalization capabilities (R2 = 0.52, MSE = 0.32) over the SVM model (R2 = 0.32, MSE = 0.45). (4) The spatial distribution of surface SOCD in the study area exhibits a decreasing gradient from west to east and from south to north. The total carbon stock in the study area is estimated at approximately 10.76 × 106t. (5) The integration of soil attribute variables, climatic variables, remote sensing data, and machine learning techniques holds significant promise for the high-precision and high-quality mapping of soil organic carbon density (SOCD) in agricultural soils.

Cold threat and moisture deficit induced individual tree mortality via 25-year monitoring in seminatural mixed forests, northeastern China

Accurately predicting tree mortality in mixed forests sets a challenge for conventional models because of large uncertainty, especially under changing climate. Machine learning algorithms had potential for predicting individual tree mortality with higher accuracy via filtering the relevant climatic and environmental factors. In this study, the sensitivity of individual tree mortality to regional climate was validated by modeling in seminatural mixed coniferous forests based on 25-year observations in northeast of China. Three advanced machine learning and deep learning algorithms were employed, including support vector machines, multi-layer perceptron, and random forests. Mortality was predicted by the effects of multiple inherent and environmental factors, including tree size and growth, topography, competition, stand structure and regional climate. All three types of models performed satisfactorily with their values of the areas under receiving operating characteristic curve (AUC) > 0.9. With tree growth, competition and regional climate as input variables, a model based on random forests showed the highest values of the explained variance score (0.862) and AUC (0.914). Since the trees were vulnerable despite their species, mortality could occur after growth limit induced by insufficient or excessive sun radiation during growing seasons, cold threat caused thermal insufficiency in winters, and annual moisture constraints in these mixed coniferous forests. Our findings could enrich basic knowledge on individual tree mortality caused by water and heat inadequacy with the negative impacts of global warming. Successful individual tree mortality modeling via advanced algorithms in mixed forests could assist in adaptive forest ecology modeling in large areas.

Benchmarking recent computational tools for DNA-binding protein identification

Identification of DNA-binding proteins (DBPs) is a crucial task in genome annotation, as it aids in understanding gene regulation, DNA replication, transcriptional control, and various cellular processes. In this paper, we conduct an unbiased benchmarking of 11 state-of-the-art computational tools as well as traditional tools such as ScanProsite, BLAST, and HMMER for identifying DBPs. We highlight the data leakage issue in conventional datasets leading to inflated performance. We introduce new evaluation datasets to support further development. Through a comprehensive evaluation pipeline, we identify potential limitations in models, feature extraction techniques, and training methods, and recommend solutions regarding these issues. We show that combining the predictions of the two best computational tools with BLAST-based prediction significantly enhances DBP identification capability. We provide this consensus method as user-friendly software. The datasets and software are available at https://github.com/Rafeed-bot/DNA_BP_Benchmarking.

Machine learning models for prediction of nutrient concentrations in surface water in an agricultural watershed

Prediction and quantification of nutrient concentrations in surface water has gained substantial attention during recent decades because excess nutrients released from agricultural and urban watersheds can significantly deteriorate surface water quality. Machine learning (ML) models are considered an effective tool for better understanding and characterization of nutrient release from agricultural fields to surface water. However, to date, no systematic investigations have examined the implementation of different classification and regression ML models in agricultural settings to predict nutrient concentrations in surface water using a group of input variables including climatological (e.g., precipitation), hydrological (e.g., stream flow) and field characteristics (i.e., land and crop use). In the current study, multiple classification (e.g., decision trees) and regression (e.g., regression trees) ML models were applied on a dataset pertaining to surface water quality in an agricultural watershed in southern Ontario, Canada (i.e., Upper Parkhill watershed). The target variables of these models were the nutrient concentrations in surface water including nitrate, total phosphorus, soluble reactive phosphorus, and total dissolved phosphorus. These target variables were predicted using physical and chemical water parameters of surface water (e.g., temperature and DO), climatological, hydrological, and field conditions as the input variables. The performance of these different models was assessed using various evaluation metrics such as classification accuracy (CA) and coefficient of determination (R2) for classification and regression models, respectively. In general, both the ensemble bagged trees and logistic regression (CA ≥ 0.72), and exponential Gaussian process regression (R2≥ 0.93) models were the optimal classification and regression ML algorithms, respectively, where they resulted in the highest prediction accuracy of the target variables. The insights and outcomes of the current study demonstrates that ML models can be employed to effectively predict and quantify the nutrient concentrations in surface waters to supplement field-collected monitoring data in agricultural watersheds, assisting in maintaining high quality of the available surface water resources.

Fusing 1H NMR and Raman experimental data for the improvement of wine recognition models

The present study proposes the development of new wine recognition models based on Artificial Intelligence (AI) applied to the mid-level data fusion of 1H NMR and Raman data. In this regard, a supervised machine learning method, namely Support Vector Machines (SVMs), was applied for classifying wine samples with respect to the cultivar, vintage, and geographical origin. Because the association between the two data sources generated an input space with a high dimensionality, a feature selection algorithm was employed to identify the most relevant discriminant markers for each wine classification criterion, before SVM modeling. The proposed data processing strategy allowed the classification of the wine sample set with accuracies up to 100% in both cross-validation and on an independent test set and highlighted the efficiency of 1H NMR and Raman data fusion as opposed to the use of a single-source data for differentiating wine concerning the cultivar and vintage.

Addressing statistical challenges in the analysis of proteomics data with extremely small sample size: a simulation study

BACKGROUND

One of the most promising approaches for early and more precise disease prediction and diagnosis is through the inclusion of proteomics data augmented with clinical data. Clinical proteomics data is often characterized by its high dimensionality and extremely limited sample size, posing a significant challenge when employing machine learning techniques for extracting only the most relevant information. Although there is a wide array of statistical techniques and numerous analysis pipelines employed in proteomics data analysis, it is unclear which of these methods produce the most efficient, reproducible, and clinically meaningful results.

RESULTS

In this study, we compared 9 unique analysis schemes comprised of different machine learning and dimensionality reduction methods for the analysis of simulated proteomics data consisting of 1317 proteins measured in 26 subjects (i.e., 13 controls and 13 cases). In scenarios where the sample size is extremely small (i.e., n < 30), all schemes resulted in an exceptionally high level of performance metrics, indicating potential overfitting. While performance metrics did not exhibit significant differences across schemes, the set of proteins selected to be discriminatory between groups demonstrated a substantial level of heterogeneity. However, despite heterogeneity in the selected proteins, their biological pathways and genetic diseases exhibited similarities. A sensitivity analysis conducted using varying sample sizes indicated that the stability of a set of selected biomarkers improves with larger sample sizes within a scheme.

CONCLUSIONS

When the aim of the study is to identify a statistical model that best distinguishes between cohort groups using proteomics data and to uncover the biological pathways and disorders common among the selected proteins, the majority of widely used analysis pipelines perform similarly. However, if the main objective is to pinpoint a set of selected proteins that wield significant influence in discriminating cohort groups and utilize them for subsequent investigations, meticulous consideration is necessary when opting for statistical models, due to the possibility of heterogeneity in the sets of selected proteins.

Assessing the Performance of Machine Learning Methods Trained on Public Health Observational Data: A Case Study From COVID-19

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

Predicting Early Treatment Effectiveness in Bell's Palsy Using Machine Learning: A Focus on Corticosteroids and Antivirals

Purpose

Facial nerve paralysis, particularly Bell's palsy, manifests as a rapid onset of unilateral facial weakness or paralysis. Despite most patients recovering within three to six months, a significant proportion experience poor recovery. This study utilized six machine learning models to investigate the effectiveness of early treatment in Bell's palsy.

Patients and Methods

We applied data from 17 hospitals in Scotland to predict treatment outcomes. Patients were randomized into four groups: Prednisolone (corticosteroids), Acyclovir (antivirals), both, and placebo. Outcomes, defined as full resolution of symptoms, were assessed using the House-Brackmann scale at 3 and 9 months post-treatment. We employed six different machine learning models to predict recovery outcomes and evaluated model performance using AUC, precision, recall, and F1-score.

Results

Among 493 patients, 72.6% recovered after three months and 89.5% after nine months. Logistic regression demonstrated the highest predictive performance for both 3-month (AUC = 0.751) and 9-month recovery (AUC = 0.720). Additionally, several models achieved Precision levels exceeding 0.9. We further employed the best-performing logistic regression for feature ranking, indicating that the patient's age and prednisolone administration are the most significant predictors of recovery.

Conclusion

The results highlight the potential of machine learning models in predicting the effectiveness of early treatment. This study conducted a comprehensive comparison of six different machine learning models, with the logistic regression showing the highest predictive performance for both 3-month and 9-month recovery. Additionally, feature ranking using logistic regression supported the importance of Prednisolone in treatment. Notably, our findings revealed the significance of age in prognosis evaluation for the first time. This suggests that future research should further develop age-specific prognostic models, enabling clinicians to tailor individualized treatment strategies more effectively. This previously unrecognized discovery provides a foundation for prognostic analysis in Bell's palsy patients.

Development of a prediction model for clinically-relevant fatigue: a multi-cancer approach

PURPOSE

Fatigue is the most prevalent symptom across cancer types. To support clinicians in providing fatigue-related supportive care, this study aims to develop and compare models predicting clinically relevant fatigue (CRF) occurring between two and three years after diagnosis, and to assess the validity of the best-performing model across diverse cancer populations.

METHODS

Patients with non-metastatic bladder, colorectal, endometrial, ovarian, or prostate cancer who completed a questionnaire within three months after diagnosis and a subsequent questionnaire between two and three years thereafter, were included. Predictor variables included clinical, socio-demographic, and patient-reported variables. The outcome was CRF (EORTC QLQC30 fatigue ≥ 39). Logistic regression using LASSO selection was compared to more advanced Machine Learning (ML) based models, including Extreme gradient boosting (XGBoost), support vector machines (SVM), and artificial neural networks (ANN). Internal-external cross-validation was conducted on the best-performing model.

RESULTS

3160 patients were included. The logistic regression model had the highest C-statistic (0.77) and balanced accuracy (0.65), both indicating good discrimination between patients with and without CRF. However, sensitivity was low across all models (0.22-0.37). Following internal-external validation, performance across cancer types was consistent (C-statistics 0.73-0.82).

CONCLUSION

Although the models' discrimination was good, the low balanced accuracy and poor calibration in the presence of CRF indicates a relatively high likelihood of underdiagnosis of future CRF. Yet, the clinical applicability of the model remains uncertain. The logistic regression performed better than the ML-based models and was robust across cohorts, suggesting an advantage of simpler models to predict CRF.

Handheld NIR Spectroscopy Combined with a Hybrid LDA-SVM Model for Fast Classification of Retail Milk

The EU market offers different types of milk, distinguished by origin, production method, processing technology, fat content, and other characteristics, which are often detailed on product labels. In this context, ensuring the authenticity of milk is crucial for maintaining standards and preventing fraud. Various food authenticity techniques have been employed to achieve this. Among them, near-infrared (NIR) spectroscopy is valued for its non-destructive and rapid analysis capabilities. This study evaluates the effectiveness of a miniaturized NIR device combined with support vector machine (SVM) algorithms and LDA feature selection to discriminate between four commercial milk types: high-quality fresh milk, milk labeled as mountain product, extended shelf-life milk, and TSG hay milk. The results indicate that NIR spectroscopy can effectively classify milk based on the type of milk, relying on different production systems and heat treatments (pasteurization). This capability was greater in distinguishing high-quality mountain and hay milk from the other types, while resulting in less successful class assignment for extended shelf-life milk. This study demonstrated the potential of portable NIR spectroscopy for real-time and cost-effective milk authentication at the retail level.

Quantitative expression of LNAPL pollutant concentrations in capillary zone by coupling multiple environmental factors based on random forest algorithm

The capillary zone plays a crucial role in migration and transformation of pollutants. Light nonaqueous liquids (LNAPLs) have become the main organic pollutant in soil and groundwater environments. However, few studies have focused on the concentration distribution characteristics and quantitative expression of LNAPL pollutants within capillary zone. In this study, we conducted a sandbox-migration experiment using diesel oil as a typical LNAPL pollutant, with the capillary zone of silty sand as the research object. The variation characteristics of LNAPL pollutants (total petroleum hydrocarbon) concentration and environmental factors (moisture content, electrical conductivity, pH, and oxidationreduction potential) were essentially consistent at different locations with the same height. These characteristics differed within range of 10.0-50.0 cm and above 60.0 cm from groundwater. A model for quantitative expression of concentrations was constructed by coupling multiple environmental factors of 968 sets-7744 data via random forest algorithm. The goodness of fit (R2) for both training and test sets was greater than 0.90, and the mean absolute percentage error (MAPE) was less than 16.00 %. The absolute values of relative errors in predicting concentrations at characteristic points were less than 15.00 %. The constructed model can accurately and quantitatively express and predict concentrations in capillary zone.

Characterizing the Contribution of Dependent Features in XAI Methods

Explainable Artificial Intelligence (XAI) provides tools to help understanding how AI models work and reach a particular decision or outcome. It helps to increase the interpretability of models and makes them more trustworthy and transparent. In this context, many XAI methods have been proposed to make black-box and complex models more digestible from a human perspective. However, one of the main issues that XAI methods have to face especially when dealing with a high number of features is the presence of multicollinearity, which casts shadows on the robustness of the XAI outcomes, such as the ranking of informative features. Most of the current XAI methods either do not consider the collinearity or assume the features are independent which, in general, is not necessarily true. Here, we propose a simple, yet useful, proxy that modifies the outcome of any XAI feature ranking method allowing to account for the dependency among the features, and to reveal their impact on the outcome. The proposed method was applied to SHAP, as an example of XAI method which assume that the features are independent. For this purpose, several models were exploited for a well-known classification task (males versus females) using nine cardiac phenotypes extracted from cardiac magnetic resonance imaging as features. Principal component analysis and biological plausibility were employed to validate the proposed method. Our results showed that the proposed proxy could lead to a more robust list of informative features compared to the original SHAP in presence of collinearity.

Enhancing random forest predictive performance for foot and mouth disease outbreaks in Uganda: a calibrated uncertainty prediction approach for varying distributions

Foot-and-mouth disease poses a significant threat to both domestic and wild cloven-hoofed animals, leading to severe economic losses and jeopardizing food security. While machine learning models have become essential for predicting foot-and-mouth disease outbreaks, their effectiveness is often compromised by distribution shifts between training and target datasets, especially in non-stationary environments. Despite the critical impact of these shifts, their implications in foot-and-mouth disease outbreak prediction have been largely overlooked. This study introduces the Calibrated Uncertainty Prediction approach, designed to enhance the performance of Random Forest models in predicting foot-and-mouth disease outbreaks across varying distributions. The Calibrated Uncertainty Prediction approach effectively addresses distribution shifts by calibrating uncertain instances for pseudo-label annotation, allowing the active learner to generalize more effectively to the target domain. By utilizing a probabilistic calibration model, Calibrated Uncertainty Prediction pseudo-annotates the most informative instances, refining the active learner iteratively and minimizing the need for human annotation and outperforming existing methods known to mitigate distribution shifts. This reduces costs, saves time, and lessens the dependence on domain experts while achieving outstanding predictive performance. The results demonstrate that Calibrated Uncertainty Prediction significantly enhances predictive performance in non-stationary environments, achieving an accuracy of 98.5%, Area Under the Curve of 0.842, recall of 0.743, precision of 0.855, and an F1 score of 0.791. These findings underscore Calibrated Uncertainty Prediction's ability to overcome the vulnerabilities of existing ML models, offering a robust solution for foot-and-mouth disease outbreak prediction and contributing to the broader field of predictive modeling in infectious disease management.

Exploring blood-brain barrier passage using atomic weighted vector and machine learning

CONTEXT

This study investigates the potential of leveraging molecular properties, as determined by MD-LOVIs software and machine learning techniques, to predict the ability of compounds to cross the blood-brain barrier (BBB). Accurate prediction of BBB permeation is critical for the development of central nervous system (CNS) drugs. The study applies various machine learning models, including both classification and regression techniques, to predict BBB passage and molecular activity. Notably, classification models such as GBM-AWV (accuracy = 0.801), GLM-CN (accuracy = 0.808), SVMPoly-means (accuracy = 0.980), SVMPoly-AC (accuracy = 0.980), SVMPoly-MI_TI_SI (accuracy = 0.900), SVMPoly-GI (accuracy = 0.900), RF-means (accuracy = 0.870), and GLM-means (accuracy = 0.818) demonstrate high accuracy in predicting BBB passage. In contrast, regression models like ES-RLM-AG (R2 = 0.902), IB-IBK (R2 = 0.82), IB-Kstar (R2 = 0.834), IB-MLP (R2 = 0.843), and DRF-AWV (R2 = 0.810) exhibit strong performance in predicting molecular activity. The results show that classification models like GBM-AWV, GLM-CN, and SVMPoly variants, as well as regression models like ES-RLM-AG and IB-MLP, achieve high performance, demonstrating the effectiveness of machine learning in predicting BBB permeability.

METHODS

The computational methods employed in this study include the MD-LOVIs software for generating molecular descriptors and several machine learning algorithms, including gradient boosting machines (GBM), generalized linear models (GLM), support vector machines (SVM) with polynomial kernels, random forests (RF), ensemble regression models, and instance-based learning algorithms. These models were trained and validated using various datasets to predict BBB passage and molecular activity, with the performance metrics reported for each model. Standard computational techniques were employed throughout, ensuring the reliability of the predictions.

Comparison of Explainable Artificial Intelligence Model and Radiologist Review Performances to Detect Breast Cancer in 752 Patients

OBJECTIVES

Breast cancer is a type of cancer caused by the uncontrolled growth of cells in the breast tissue. In a few cases, erroneous diagnosis of breast cancer by specialists and unnecessary biopsies can lead to various negative consequences. In some cases, radiologic examinations or clinical findings may raise the suspicion of breast cancer, but subsequent detailed evaluations may not confirm cancer. In addition to causing unnecessary anxiety and stress to patients, such diagnosis can also lead to unnecessary biopsy procedures, which are painful, expensive, and prone to misdiagnosis. Therefore, there is a need for the development of more accurate and reliable methods for breast cancer diagnosis.

METHODS

In this study, we proposed an artificial intelligence (AI)-based method for automatically classifying breast solid mass lesions as benign vs malignant. In this study, a new breast cancer dataset (Breast-XD) was created with 791 solid mass lesions belonging to 752 different patients aged 18 to 85 years, which were examined by experienced radiologists between 2017 and 2022.

RESULTS

Six classifiers, support vector machine (SVM), K-nearest neighbor (K-NN), random forest (RF), decision tree (DT), logistic regression (LR), and XGBoost, were trained on the training samples of the Breast-XD dataset. Then, each classifier made predictions on 159 test data that it had not seen before. The highest classification result was obtained using the explainable XGBoost model (X2GAI) with an accuracy of 94.34%. An explainable structure is also implemented to build the reliability of the developed model.

CONCLUSIONS

The results obtained by radiologists and the X2GAI model were compared according to the diagnosis obtained from the biopsy. It was observed that our developed model performed well in cases where experienced radiologists gave false positive results.

Knowledge graph based reasoning in medical image analysis: A scoping review

Automated computer-aided diagnosis (CAD) is becoming more significant in the field of medicine due to advancements in computer hardware performance and the progress of artificial intelligence. The knowledge graph is a structure for visually representing knowledge facts. In the last decade, a large body of work based on knowledge graphs has effectively improved the organization and interpretability of large-scale complex knowledge. Introducing knowledge graph inference into CAD is a research direction with significant potential. In this review, we briefly review the basic principles and application methods of knowledge graphs firstly. Then, we systematically organize and analyze the research and application of knowledge graphs in medical imaging-assisted diagnosis. We also summarize the shortcomings of the current research, such as medical data barriers and deficiencies, low utilization of multimodal information, and weak interpretability. Finally, we propose future research directions with possibilities and potentials to address the shortcomings of current approaches.

Physiological model-based machine learning for classifying patients with binge-eating disorder (BED) from the Oral Glucose Tolerance Test (OGTT) curve

BACKGROUND AND OBJECTIVE

Binge eating disorder (BED) is the most frequent eating disorder, often confused with obesity, with which it shares several characteristics. Early identification could enable targeted therapeutic interventions. In this study, we propose a hybrid pipeline that, starting from plasma glucose data acquired during the Oral Glucose Tolerance Test (OGTT), allows us to classify the two types of patients through computational modeling and artificial intelligence.

METHODS

The proposed hybrid pipeline integrates a classical mechanistic model of delayed differential equations (DDE) that describes glucose-insulin dynamics with machine learning (ML) methods. Ad hoc techniques, including structural identifiability analysis, have been employed for refining and evaluating the mathematical model. Additionally, a dedicated pipeline for identifying and optimizing model parameters has been applied to obtain reliable estimates. Robust feature extraction and classifier selection processes were developed to ensure the optimal choice of the best-performing classifier.

RESULTS

By leveraging parameters estimated from the mechanistic model alongside easily obtainable patient information (such as glucose levels at 30 and 120 min post-OGTT, glycated hemoglobin (Hb1Ac), body mass index (BMI), and waist circumference), our approach facilitates accurate classification of patients, enabling tailored therapeutic interventions.

CONCLUSION

Initial findings, focusing on correctly categorizing patients with BED based on metabolic data, demonstrate promising outcomes. These results suggest significant potential for refinement, including exploration of alternative mechanistic models and machine learning algorithms, to enhance classification accuracy and therapeutic strategies.