False Positive Research Articles

Implementation of machine learning algorithms to screen for advanced liver fibrosis in metabolic dysfunction-associated steatotic liver disease (MASLD): an in-depth explanatory analysis.

Background This study aimed to train machine learning algorithms(MLAs) to detect advanced fibrosis(AF) in MASLD patients at the level of primary care setting and to explain the predictions to ensure responsible use by clinicians. Methods Readily available features of 618 MASLD patients followed up at a tertiary center were used to train five MLAs. AF was defined as liver stiffness≥9.3 kPa, measured via 2-dimension shear wave elastography(n=495) or liver biopsy≥F3(n=123). MLAs were compared to Fibrosis-4 index(FIB-4) and NAFLD fibrosis score(NFS) on 540 MASLD patients from the primary care setting as validation. Feature importance, partial dependence, and shapely additive explanations(SHAP) were utilized for explanation. Results Extreme gradient boosting(XGBoost) achieved an AUC=0.91,outperforming FIB-4(AUC=0.78) and NFS(AUC=0.81, both p<0.05) with specificity=76% vs. 59% and 48% for FIB-4≥1.3 and NFS≥-1.45, respectively(p<0.05). Its sensitivity(91%) was superior to FIB-4(79%). XGBoost confidently excluded AF (negative predictive value=99%) with the highest positive predictive value (31%), superior to FIB-4 and NFS (all p<0.05). The most important features were HbA1c and GGT with a steep increase in AF probability at HbA1c>6.5%. The strongest interaction was between AST and age. XGBoost, but not logistic regression, extracted informative patterns from ALT, LDL-c,and ALP(p<0.001). One quarter of the false positives (FP) were correctly reclassified with only one additional false negative based on the SHAP values of GGT, platelets, and ALT which were found to be associated with a FP classification. Conclusions: An explainable XGBoost algorithm was demonstrated superior to FIB-4 and NFS for screening of AF in MASLD patients at the primary care setting. The algorithm also proved safe for use as clinicians can understand the predictions and flag FP classifications.

Bayesian intelligence for medical diagnosis: a pilot study on patient disposition for emergency medicine chest pain.

Clinicians can rapidly and accurately diagnose disease, learn from experience, and explain their reasoning. Computational Bayesian medical decision-making might replicate this expertise. This paper assesses a computer system for diagnosing cardiac chest pain in the emergency department (ED) that decides whether to admit or discharge a patient. The system can learn likelihood functions by counting data frequency. The computer compares patient and disease data profiles using likelihood. It calculates a Bayesian probabilistic diagnosis and explains its reasoning. A utility function applies the probabilistic diagnosis to produce a numerical BAYES score for making a medical decision. We conducted a pilot study to assess BAYES efficacy in ED chest pain patient disposition. Binary BAYES decisions eliminated patient observation. We compared BAYES to the HEART score. On 100 patients, BAYES reduced HEART's false positive rate 18-fold from 58.7 to 3.3 %, and improved ROC AUC accuracy from 0.928 to 1.0. The pilot study results were encouraging. The data-driven BAYES score approach could learn from frequency counting, make fast and accurate decisions, and explain its reasoning. The computer replicated these aspects of diagnostic expertise. More research is needed to reproduce and extend these finding to larger diverse patient populations.

SLICK: Strong Lensing Identification of Candidates Kindred in gravitational wave data

ABSTRACT By the end of the next decade, we hope to have detected strongly lensed gravitational waves by galaxies or clusters. Although there exist optimal methods for identifying lensed signal, it is shown that machine learning (ML) algorithms can give comparable performance but are orders of magnitude faster than non-ML methods. We present the SLICK pipeline which comprises a parallel network based on deep learning. We analyse the Q-transform maps (QT maps) and the Sine-Gaussian projection (SGP) maps generated for the binary black hole signals injected in Gaussian as well as real noise. We compare our network performance with the previous work and find that the efficiency of our model is higher by a factor of 5 at a false positive rate of 0.001. Further, we show that including SGP maps with QT maps data result in a better performance than analysing QT maps alone. When combined with sky localization constraints, we hope to get unprecedented accuracy in the predictions than previously possible. We also evaluate our model on the real events detected by the LIGO–Virgo collaboration and find that, at a threshold of 0.75 our network correctly classifies all of them, consistent with non-detection of lensing.

Mitigating Maritime Cybersecurity Risks Using AI-Based Intrusion Detection Systems and Network Automation During Extreme Environmental Conditions

The maritime industry is increasingly confronted with a myriad of cybersecurity challenges exacerbated by extreme environmental conditions, technological advancements, and heightened reliance on automation. This review paper discusses the intersection of these factors, focusing on the adoption of artificial intelligence (AI)-based intrusion detection systems (IDS) and network automation as vital strategies for mitigating cybersecurity risks. The paper begins by outlining the unique cybersecurity threats faced by the maritime sector, which include data breaches, phishing attacks, and malware threats, all amplified by adverse weather and geographical isolation. In light of these challenges, the rationale for integrating AI-driven solutions into maritime operations is discussed. AI-based IDS can enhance threat detection capabilities through advanced machine learning algorithms that adapt to evolving cyber threats while minimizing false positives. Additionally, network automation can improve connectivity and data security, facilitating real-time monitoring and response to incidents. The review also addresses the critical need for collaboration between maritime and technology industries, emphasizing how partnerships can foster innovation and provide tailored solutions to the sector’s specific needs. Furthermore, the paper examines current implementations and case studies that illustrate successful applications of AI and automation in adverse maritime conditions. While recognizing the potential benefits, the review highlights the technical and operational challenges inherent in these implementations, including data integration, regulatory compliance, and cultural differences between sectors. Ultimately, this paper aims to provide a comprehensive overview of the state of maritime cybersecurity and the pivotal role of AI and automation in shaping a resilient, secure maritime future. The findings underscore the importance of ongoing research and development, collaborative efforts, and the necessity of adaptable strategies to safeguard maritime operations against the evolving landscape of cyber threats.

Diagnostic Performances of 18F-Fluorocholine PET/CT as First-Line Functional Imaging Method for Localization of Hyperfunctioning Parathyroid Tissue in Primary Hyperparathyroidism

Rationale and ObjectivesThis study evaluated the diagnostic performance of 18F-fluorocholine (FCH) PET/CT as the first-line functional imaging method for preoperative localization of hyperfunctioning parathyroid glands (HPGs) in patients with primary hyperparathyroidism (PHPT). Materials and MethodsThis retrospective single-centre study included 80 consecutive patients with PHPT, referred for FCH PET/CT between January 2018 and July 2022, and who subsequently underwent surgery. The diagnostic performance of FCH PET/CT was compared to histological results for per-lesion analysis, and to postoperative resolution of biochemical PHPT for per-patient analysis. Results18F-FCH-PET/CT revealed 95 positive foci in 77/80 patients and was negative in 3/80 patients. Postoperative resolution of HPT was obtained in 67/80 patients (84%). Per-lesion analysis showed 80 true positives, 5 true negatives, 11 false negatives and 8 false positives. Seven PET positive foci could not be compared with histology. In a first per-lesion analysis, excluding these 7 anomalies, sensitivity and positive predictive value (PPV) of FCH PET/CT were 88% (95% CI: 79-94) and 91% (95% CI: 87-94), respectively. In a second per-lesion analysis considering the 7 anomalies as false positives (maximum bias analysis), PPV was 84% (95% CI: 80%-87%). By per-patient analysis, FCH PET/CT correctly identified and located all pathological glands in 56/80 (70%, 95% CI: 59-80) patients. Conclusion18F-Fluorocholine PET/CT appears to be an effective pre-surgical imaging method for localization of hyperfunctioning parathyroid tissue in patients with PHPT.

An Improved Design for a Cloud Intrusion Detection System Using Hybrid Features Selection Approach with ML Classifier

The rapid adoption of cloud computing has introduced new security concerns, particularly regarding data privacy and protection from sophisticated cyberattacks. Intrusion Detection Systems (IDSs) based on Machine Learning (ML) offer improved detection of malicious behaviour in network traffic. However, existing IDS solutions struggle with large-dimensional data and imbalanced datasets, resulting in higher false positive rates (FPR) and reduced detection rates (DR). In this paper, we propose an improved Cloud IDS leveraging a hybrid features selection approach using Information Gain (IG), Chi-square (CS), and Particle Swarm Optimization (PSO). To address the challenge of imbalanced datasets, the Synthetic Minority Over-sampling Technique (SMOTE) is employed. The Random Forest (RF) classifier is used to detect and classify attacks. The system is evaluated on the UNSW-NB15 and Kyoto datasets, achieving accuracies of over 98% and 99%, respectively, outperforming other methods in terms of detection accuracy.

Identifications of False Positives Amongst Sodium(I) Cations in Protein Three-Dimensional Structures—A Validation Approach Extendible to Any Alkali or Alkaline Earth Cation and to Any Monoatomic Anion

Validation of the data deposited in the Protein Data Bank is of the upmost importance, since many other databases, data mining processes, and artificial intelligence tools are strictly grounded on them. The present paper is divided into two parts. The first part describes and analyzes validation methods that have been designed and used by the structural biology community. Everything began with the Ramachandran plot, with its allowed and disallowed types of backbone conformations, and evolved in different directions, with the inclusion of additional stereochemical features, distributions’ analyses of structural moieties, and scrutiny of structure factor amplitudes across the reciprocal lattice. The second part of the paper is focused on the largely unexplored problem of the high number of false positives amongst the sodium(I) cations observed in protein crystal structures. It is demonstrated that these false positives, which are atoms wrongly identified with sodium, can be identified by using electrostatic considerations and it is anticipated that this approach can be extended to other alkali and alkaline earth cations or to monoatomic anions. In the end, I think a global initiative, accessible to all volunteers and possibly overseen by the Protein Data Bank, should take the place of the numerous web servers and software applications by providing the community with a select few reliable and widely accepted tools.

Simulation-based evaluation of advanced threat detection and response in financial industry networks using zero trust and blockchain technology

The financial sector is increasingly facing advanced cyber threats, necessitating a shift from traditional security measures to more dynamic frameworks. This study presents a novel integration of Zero Trust architecture with hybrid access control system and blockchain technology to enhance security in financial institutions. Zero Trust enforces continuous authentication and dynamic access controls, while blockchain secures digital identities and transaction logs through its immutable ledger, ensuring data integrity and non-repudiation. The proposed framework, evaluated using OMNeT++ simulations enhanced by Ethereum-Ganache, shows improved detection accuracy, reduced false positives, and increased resistance to insider threats and other attacks. It also strengthens compliance with regulatory requirements through robust audit trails, providing enhanced protection for sensitive financial data.

CTARR: A fast and robust method for identifying anatomical regions on CT images via atlas registration

Medical image analysis tasks often focus on regions or structures located in a particular location within the patient’s body. Often large parts of the image may not be of interest for the image analysis task. When using deep-learning based approaches, this causes an unnecessary increases the computational burden during inference and raises the chance of errors. In this paper, we introduce CTARR, a novel generic method for CT Anatomical Re- gion Recognition. The method serves as a pre-processing step for any deep learning-based CT image analysis pipeline by automatically identifying the pre-defined anatomical region that is relevant for the follow-up task and removing the rest. It can be used in (i) image segmentation to prevent false positives in anatomically implausible regions and speeding up the inference, (ii) image classification to produce image crops that are consistent in their anatomical context, and (iii) image registration by serving as a fast pre-registration step. Our proposed method is based on atlas registration and provides a fast and robust way to crop any anatomical region encoded as one or multiple bounding box(es) from any unlabeled CT scan of the brain, chest, abdomen and/or pelvis. We demonstrate the utility and robustness of the proposed method in the context of medical image segmentation by evaluating it on six datasets of public segmentation challenges. The foreground voxels in the regions of interest are preserved in the vast majority of cases and tasks (97.45-100%) while taking only fractions of a seconds to compute (0.1-0.21s) on a deep learning worksta- tion and greatly reducing the segmentation runtime (2.0-12.7x). Our code is available at <a href='https://github.com/ThomasBudd/ctarr'>https://github.com/ThomasBudd/ctarr</a>

Evaluating Bioinformatics Processing of Somatic Variant Detection in cfDNA Using Targeted Sequencing with UMIs.

Circulating tumor DNA (ctDNA) is a promising cancer biomarker, but accurately detecting tumor mutations in cell-free DNA (cfDNA) is challenging due to their low frequency and sequencing errors. Our study benchmarked Mutect2, VarScan2, shearwater, and DREAMS-vc using deep targeted sequencing of cfDNA with Unique Molecular Identifiers (UMIs) from 111 colorectal cancer patients. Performance was assessed at both the mutation level (distinguish tumor variants from errors) and the sample level (detect if an individual has cancer). Additionally, we investigated the effects of various UMI grouping and consensus strategies. The shearwater-AND variant calling method demonstrated the highest precision in detecting tumor-derived mutations from plasma, and reached the highest ROC-AUC of 0.984 for sample classification in tumor-informed cfDNA analyses. DREAMS-vc exhibited the highest ROC-AUC of 0.808 for sample classification in tumor-agnostic studies. We also found that sequencing depth differences in PBMCs could lead to false positives, particularly with VarScan2 and Mutect2, which was addressed by downsampling to equivalent mean depths. Additionally, network-based UMI grouping methods outperformed those using identical UMIs when all reads were retained. Our findings emphasize that the optimal variant caller depends on the study context-whether focused on mutation or sample classification, and whether conducted under tumor-informed or tumor-agnostic conditions.

Optimizing Machine Learning-Based Intrusion Detection Systems for Enhanced Security in Cloud Computing Environments.

In today’s cloud computing environments, robust network security is crucial due to the inherently distributed and dynamic nature of cloud systems. Traditional Network Intrusion Detection Systems (NIDS) often face challenges in handling large-scale, rapidly evolving data and sophisticated attack vectors unique to cloud settings. This paper presents an optimized Machine Learning (ML)-based NIDS framework designed specifically for cloud infrastructures, leveraging feature selection, dimensionality reduction, and algorithmic tuning techniques to enhance detection accuracy and reduce false positives. Integrating supervised and unsupervised learning methods, this NIDS system can detect both known and novel attack types efficiently. By utilizing cloud-specific datasets and real-time traffic monitoring, it adapts well to the scalability and elasticity requirements of cloud environments. Additionally, the system incorporates automated optimization techniques, such as hyperparameter tuning and ensemble learning, to further enhance performance. Experimental results show improvements in detection rates, reduced computational overhead, and better scalability, making it a promising solution for modern cloud-based infrastructures.

15 Predictive validity of two developmental screening tools at the 18-month visit and neurodevelopmental consultation at an average age of 8 years

Abstract Background The Canadian Paediatric Society (CPS) recommends an enhanced 18-month well-baby visit incorporating a developmental screening tool to stimulate discussion with parents, and notes that a widely used tool is the Nipissing District Developmental Screen (NDDS). For early detection of autism spectrum disorder, the CPS notes that a commonly used screening tool is the Infant Toddler Checklist (ITC). Objectives To examine the predictive validity of the ITC and NDDS at the 18-month visit using later neurodevelopmental consultation as the criterion measure; and examine whether positive developmental screening was associated with health care utilization (HCU). Design/Methods Using a prospective design, parents completed the ITC and the NDDS at the 18-month visit in primary care practices. Seven types of HCU were collected from health administrative databases, via linkage using the child’s health insurance number, including: neurodevelopmental consultation, special paediatric consultation, scheduled primary care visit billings, unscheduled primary care or minor visit billings, other non-primary care visit billings, hospitalizations, and emergency department visits. Covariates included child and family characteristics. We calculated screening test properties (using neurodevelopmental consultation as the criterion) and used multivariable negative binomial regression to estimate rate ratios (RR) for each of the 7 types of HCU. Results For both the ITC (n=1460) and NDDS (n=802) cohorts, mean age at screening was 18 months and mean age at follow up was 8 years. Of the ITC cohort, 11% were screen positive, 2.6% had a neurodevelopmental consultation. A positive ITC had 40% sensitivity (95% CI 24%-57%), 90% specificity (95% CI 88%-91%), 10% false positive rate (95% CI 9%-12%). A positive ITC was associated with 6 of 7 types of HCU, including neurodevelopmental consultation (RR=2.78, 95% CI 1.37-5.67). Of the NDDS cohort, 36% were screen positive, 2.5% had a neurodevelopmental consultation. A positive NDDS had a sensitivity of 50% (95% CI 27%-73%), specificity of 65% (95% CI 61%-68%), false positive rate of 35% (95% CI 31%-40%). A positive NDDS was associated with only 1 of 7 types of HCU: unscheduled primary care or minor visit billings 1.24 (95% CI 1.01, 1.53, p=0.04). Conclusion The predictive validity of the ITC was stronger than the NDDS due to high specificity and low false positive rates. The low sensitivity of the ITC is similar to other screening tools used at 18 months, highlighting the importance of ongoing developmental surveillance. The false positive rate of the NDDS is unacceptably high for a screening tool.

Machine Learning-Driven Data Valuation for Optimizing High-Throughput Screening Pipelines.

In the rapidly evolving field of drug discovery, high-throughput screening (HTS) is essential for identifying bioactive compounds. This study introduces a novel application of data valuation, a concept for evaluating the importance of data points based on their impact, to enhance drug discovery pipelines. Our approach improves active learning for compound library screening, robustly identifies true and false positives in HTS data, and identifies important inactive samples in an imbalanced HTS training, all while accounting for computational efficiency. We demonstrate that importance-based methods enable more effective batch screening, reducing the need for extensive HTS. Machine learning models accurately differentiate true biological activity from assay artifacts, streamlining the drug discovery process. Additionally, importance undersampling aids in HTS data set balancing, improving machine learning performance without omitting crucial inactive samples. These advancements could significantly enhance the efficiency and accuracy of drug development.

A Supervised Ml Algorithm for Detecting and Predicting Fraud Credit Card Transactions

A Credit card fraud presents an escalating threat in today's digital economy, leading to significant financial losses for consumers, businesses, and financial institutions. Traditional detection methods are increasingly inadequate due to the rise in online transactions and the evolving complexity of fraudulent activities. This research aims to create a reliable supervised machine learning model designed to detect and predict fraudulent credit card transactions in real-time. Our approach leverages advanced algorithms and extensive datasets to enhance transaction security, thereby mitigating financial risks. Current systems, like FICO's Fraud Detection System (FDS), utilize a range of supervised learning techniques, including Decision Trees, Random Forests, and Neural Networks. However, they face limitations, including data quality issues, high false positive rates, and the need for continuous retraining to adapt to evolving fraud patterns. To address these challenges, we propose an innovative anomaly detection method based on the Isolation Forest algorithm. Unlike traditional methods that rely on distance and density measures, Isolation Forests isolate anomalies by randomly selecting features and split values, making it more efficient in identifying fraudulent transactions, especially in imbalanced datasets. The proposed system boasts low linear time complexity and minimal memory requirements, enabling the construction of high-performing models even with large datasets. The research demonstrates that the Isolation Forest approach significantly outperforms traditional models in detecting credit card fraud, offering a promising solution for enhancing security in the financial ecosystem.

PsyneuroNet architecture for multi-class prediction of neurological disorders

Early detection and diagnosis of psychiatric disorders are crucial for timely intervention, improved treatment outcomes, and preventing complications. The purpose of this study is to use artificial intelligence methods to diagnose symptoms of psychiatric conditions such as depression, schizophrenia, sleep disorder, and Alzheimer’s disease in epileptic patients using EEG data. The research uses a cross-validation procedure with ten iterations and looks at three data sets: the CHB-MIT dataset with EEG recordings from neonatal subjects, the DEAP dataset containing physiological signals and emotional ratings, and the OpenNeuro dataset containing multichannel EEG. Preprocessing involves dividing the EEG signals into five brain wave frequency components and applying filtering techniques for further analysis. We introduce a novel architecture called PsyneuroNet, which aims to improve feature extraction and prediction accuracy for real-time psychiatric disorder recognition and integrates three key components: cross-normalization across convolution, a bidirectional long-short-term memory network (Bi-LSTM), and an enhanced intuition-based convolutional neural network (ICNN). The proposed methodology is evaluated based on the 99.65% prediction accuracy achieved on the CHB-MIT dataset, the 97.27% prediction accuracy achieved on the DEAP dataset, and the 98.08% prediction accuracy achieved on the OpenNeuro dataset for prediction problems. The proposed method outperforms the state-of-the-art approaches in terms of accuracy by 3.64%, with a sensitivity of 99.65%, a specificity of 99.75%, and a very low false positive rate per hour (FPR/h) of 0.3%, all achieved by using ten folds cross-validation. The PsyneuroNet’s capability to classify a diverse range of neurological disorders within a single architecture is noteworthy. This versatility makes the model suitable for addressing a broad spectrum of clinical scenarios. The model’s ability to generalize effectively to unseen data, as demonstrated by testing its performance on the DEAP and OpenNeuro datasets after training on CHB-MIT data, is a significant aspect of novelty.

Intra-operative neuromonitoring in paediatric spinal deformity surgery: a retrospective single-centre experience.

Intra-operative neuromonitoring including somatosensory evoked potentials, motor evoked potentials, and electromyography, have replaced the Stagnara wake-up test to allow early detection of neurological change during paediatric spinal deformity surgery. It is important for surgeons to recognize alerts triggered by loss of these potentials and act accordingly to prevent iatrogenic neurological damage intra-operatively. This study was conducted to determine the sensitivity and specificity of neuromonitoring alerts in paediatric spinal deformity correction surgery. A retrospective single-centre study of all patients undergoing spinal deformity surgery at a tertiary paediatric centre between 1 January 2017 and 31 December 2020 (inclusive) was conducted. Neuromonitoring alerts were identified through neurophysiology documentation, and these were correlated with neurological deficits documented in the patient record post-operatively. A total of 399 operations were included in the study, with 147 (35.7%) of these having a motor, or motor and sensory alert triggered. Fifteen (10.2% of alerts) of these patients had a post-operative neurological deficit, compared to seven (2.8% of no alerts) of those that had no neuromonitoring alert. The sensitivity for post-operative neurological deficits not resolving within 3 days was 100%, and the specificity was 65.5%. Intra-operative neuromonitoring is highly sensitive to post-operative neurological deficits lasting longer than 3 days. However, there is still scope for optimization of specificity, with many false positives identified.

Instance segmentation models for detecting floating macroplastic debris from river surface images

Quantifying the transport of floating macroplastic debris (FMPD) in waterways is essential for understanding the plastic emission from land. However, no robust tool has been developed to monitor FMPD. Here, to detect FMPD on river surfaces, we developed five instance segmentation models based on state-of-the-art You Only Look Once (YOLOv8) architecture using 7,356 training images collected via fixed-camera monitoring of seven rivers. Our models could detect FMPD using object detection and image segmentation approaches with accuracies similar to those of the pretrained YOLOv8 model. Our model performances were tested using 3,802 images generated from 107 frames obtained by a novel camera system embedded in an ultrasonic water level gauge (WLGCAM) installed in three rivers. Interestingly, the model with intermediate weight parameters most accurately detected FMPD, whereas the model with the most parameters exhibited poor performance due to overfitting. Additionally, we assessed the dependence of the detection performance on the ground sampling distance (GSD) and found that a smaller GSD for image segmentation approach and larger GSD for object detection approach are capable of accurately detecting FMPD. Based on the results from our study, more appropriate category selections need to be determined to improve the model performance and reduce the number of false positives. Our study can aid in the development of guidelines for monitoring FMPD and the establishment of an algorithm for quantifying the transport of FMPD.

Detection of circulating tumor cells in non-metastatic prostate cancer through integration of a microfluidic CTC enrichment system and multiparametric flow cytometry.

Prostate cancer (PCa) is the second most common cancer among men and the fifth leading cause of cancer death. Circulating tumor cell (CTC) enumeration and characterisation in PCa has been shown to provide valuable information on prognosis of disease, therapy management and detection of resistance. Here, Cellsway's microfluidic platform for high-throughput enrichment of intact CTC populations was used to isolate CTCs from the blood of 20 localised PCa patients and 10 healthy donor samples to evaluate the clinical performance of the technology. To enumerate and characterise CTCs, a multi-parameter flow cytometry analysis was performed on the enriched CTC suspension using CTC-specific biomarkers. CTCs were detected in 17 of 20 patient samples, which corresponds to 85% CTC positivity. The median CTC count per 7.5 ml blood was 2 (1-9). In 80% of patients (n = 16), the number of CTCs ranged from 1 to 5, and in 5% of patients (n = 1) the number of CTCs was above 5. No CTCs were observed in the blood samples of 10 healthy volunteers, demonstrating the high specificity and low risk of false positives of the technology.

EGMA: Ensemble Learning-Based Hybrid Model Approach for Spam Detection

Spam messages have emerged as a significant issue in digital communication, adversely affecting users’ mental health, personal safety, and network resources. Traditional spam detection methods often suffer from low detection rates and high false positives, underscoring the need for more effective solutions. This paper proposes the EGMA model, an ensemble learning-based hybrid approach for spam detection in SMS messages, which integrates gated recurrent unit (GRU), multilayer perceptron (MLP), and hybrid autoencoder models utilizing a majority voting algorithm. The EGMA model enhances performance by incorporating additional statistical features extracted from message content and employing text vectorization techniques, such as Term Frequency–Inverse Document Frequency (TF-IDF) and CountVectorizer. The proposed model achieved impressive classification accuracies of 99.28% on the SMS Spam Collection dataset, 99.24% on the Email Spam dataset, 99.00% on the Enron-Spam dataset, 98.71% on the Super SMS dataset, and 95.09% on UtkMl’s Twitter Spam dataset. These results demonstrate that the EGMA model outperforms individual models and existing methods in the literature, providing a robust solution for enhancing spam detection performance and effectively mitigating the threats that spam messages pose in digital communication.

Risk profiles for rheumatoid arthritis-associated interstitial lung disease in a cohort of patients with five-year follow-up

Objectives Early identification of interstitial lung disease (ILD) among patients with rheumatoid arthritis (RA) is a challenge for clinicians. The aim of this study was to evaluate screening algorithms for ILD by comparing the proportion of patients assigned a high-risk profile by three recently proposed models. Method We used the four-factor risk score, categorizing patients into high and low risk; the ILD screening criteria, categorizing patients into high, intermediate, and low risk; and the risk score for detection of subclinical RA-ILD, with four different risk categories, on patients with RA followed for 5 years after the RA diagnosis with pulmonary function tests, dyspnoea score, and pulmonary imaging. Results The four-factor risk score identified 22% of the cohort (25/115) as eligible for further ILD investigations, while the ILD screening criteria identified 37% as high risk (43/115) and 34% as intermediate risk (39/115). The risk score for detection of subclinical RA-ILD identified 44% of the cohort as being at increased risk, with 7% in the highest risk group. The agreement between high-risk groups in the two clinical ILD screening models was moderate (kappa 0.43). Three patients in the cohort had clinical or subclinical ILD, and they were identified as high risk in the two clinical models. Conclusion The three algorithms identified approximately one-third of the cohort as being at increased risk of ILD. Further development and validation of these algorithms are needed to reduce false positives and balance the potential benefit of earlier ILD diagnosis and healthcare resources used for respiratory assessment.