Search Results (338)

Background: Internet of Things (IoT) deployments face complex network dependencies and persistent data limitations that constrain machine learning (ML) intrusion detection systems (IDS). Objective: To synthesise peer-reviewed machine learning IDS research for IoT, focusing on dependency-driven failure propagation and chronic data scarcity, positioned against 2024–2025 EU regulatory requirements (NIS2, the Cyber Resilience Act). Eligibility criteria: Peer-reviewed empirical studies proposing or evaluating a machine learning or deep learning IoT intrusion detection method published in English from January 2018 (limited pre-2018 exceptions for seminal works). Information sources: IEEE Xplore, SpringerLink, Elsevier ScienceDirect, Scopus, Web of Science, and Google Scholar, searched on 12 February 2025. Risk of bias: Each candidate was scored against four criteria (objectives clarity, methodological soundness, reproducibility, IoT-security relevance); studies scoring at least 3 out of 4 were retained. Screening and scoring were performed by one reviewer, with a second reviewer independently checking 20 percent of records. Synthesis methods: Narrative thematic synthesis; heterogeneous metrics and incompatible datasets across studies precluded quantitative meta-analysis. Included studies: Of 427 records identified, 52 studies initially met inclusion criteria; a post hoc independently validated reconstruction of individual QA1–QA4 scores subsequently found that six did not meet the threshold or topical eligibility criteria, yielding a final 46-study corpus. Main findings: Generative adversarial networks (GANs) dominate dataset augmentation work, graph-based communication analysis addresses dependency modelling, and methods based on transformers or federated learning emerge from 2023 onward. Certainty of evidence: No formal grading (GRADE) applies to this narrative synthesis. Confidence in the corpus composition is high, following independent QA1–QA4 validation, while confidence in the thematic findings is moderate given single-reviewer screening and judgment-based classification. Conclusions: We identify three recurring gaps: real-time detection under resource constraints, dependency-aware detection, and regulatory compliance. Closing these gaps requires detection methods that treat IoT security as a networked and regulated system rather than an isolated device classification problem. Registration: Open Science Framework, 10.17605/OSF.IO/NMAK4 (registered retrospectively). No external funding supported this review. Full article

Maritime object detection is essential for port-based surveillance, ship tracking, and maritime security. However, practical implementation encounters two primary challenges: significant class imbalance (IMO identification numbers represent only 1.2% of labelled objects) and spatial inconsistency (predicted IMO numbers often appear beyond ship boundaries). Standard detectors and conventional class-balancing strategies fail to adequately address these difficulties, resulting in a persistent mismatch between research validation and practical performance. We present a hybrid autoencoder–YOLO framework with variable spatial regularisation. To the best of our understanding, this is the first methodology to simultaneously address class imbalance and spatial inconsistency in maritime IMO detection via reconstruction-guided learning and differentiable spatial regularisation. The model uses a YOLOv8 encoder shared by both a detection head (designed for ships and IMO numbers), and an auxiliary reconstruction decoder (a SkipDecoder with U-Net-style skip connections). A unique spatial limitation loss provides the physical restriction that each IMO number must be within a ship’s bounding box, using only anticipated boxes. Training occurs in two phases: autoencoder pretraining on the marine dataset, followed by phased joint optimisation with a curriculum schedule for the spatial weighting. In a dataset of 297 annotated images (utilising five-fold cross-validation with a 47-image preserved test set), our comprehensive model achieved 50.1% IMO AP50-95, 97.8% IMO precision, and 94.6% IMO F1-score on the test set, beating the baseline YOLOv8s by +7.8 percentage points, +6.5 percentage points, and +5.2 percentage points, respectively. Ablation studies indicate that reconstruction instruction improves IMO AP50-95 by +4.5 percentage points, while spatial regularisation adds +3.3 percentage points. Although ship detection results in a small compromise (ship AP50-95 decreases from 77.0% to 55.6%), this appears to be practically acceptable given the essential role of IMO numbers as unique ship IDs. Significantly, inference speed improves by 20% (8.0 ms per image on an NVIDIA A100 GPU) relative to YOLOv8s (10.0 ms). Comparisons with leading detectors (RetinaNet, Faster R-CNN, DETR, EfficientDet), all initialised with standard COCO-pretrained backbones and fine-tuned on our dataset, reveal that none achieve an IMO AP50-95 exceeding 42.3%, highlighting the task’s challenge and the accuracy of our design. The proposed framework shows potential to address the gap in practical implementation through reconstruction-based feature learning accompanied by a specified geometric baseline. It is precise, accurate, and fast, aligns with specific physics, and shows promise for real-time maritime surveillance applications, though we acknowledge the need for additional thorough verification across many operational environments. Full article

With the rapid proliferation of Internet of Things (IoT) technology today, the number of devices connected to each other through these networks is increasing significantly, and the security risks these devices face from cyberattacks have emerged as a clear problem. Intrusion detection systems (IDSs) are emerging as a solution to this problem and play a significant role in securing IoT-based networks. Although machine learning and deep learning-based IDS approaches have achieved high accuracy rates in detecting attacks in recent years, the lack of transparency in these models’ decision-making processes poses a major drawback in terms of reliability and explainability. To ensure that the decisions of IDS models are understandable and to address this issue, Explainable AI (XAI) approaches are being implemented. This study provides a detailed review of the current literature on XAI-enabled IDSs developed for IoT environments. The studies examined are systematically evaluated in terms of the deep learning models used, lightweight model designs, explainability methods and validations, approaches to protecting data privacy, and datasets. The literature review highlights that research is not only focused on the accuracy of attack detection but also aims to design IDS solutions that are lightweight, explainable, and privacy-preserving. This paper aims to provide a foundational resource to guide future research on reliable, explainable, and practical IoT intrusion detection systems by identifying the problems addressed in current research and highlighting the limitations in the literature. In addition, this work proposes and applies a study assessment framework to systematically assess methodological adequacy, experimental reproducibility, lightweight deployment, XAI techniques, and XAI validation. Full article

The use of Artificial Intelligence (AI) and Machine Learning (ML) in cybersecurity, especially for creating Intrusion Detection Systems (IDSs), has become increasingly important. These systems are essential for detecting malicious behaviour, identifying network issues, and stopping cyberattacks in real time. Despite extensive research on various ML and Deep Learning (DL) models for IDS, the current literature remains incomplete. It has many different datasets, methods, and evaluation standards. As cyber threats become more advanced, it is crucial to conduct a thorough analysis of ML techniques for intrusion detection. The goal of this Systematic Literature Review (SLR) is to provide a full picture of the most recent academic articles on ML-based IDS. The study addresses important research questions about the most widely used algorithms, the types of attacks and network environments covered, the methodological problems that remain unsolved, and the new trends that should shape future research. Following the PRISMA framework, we conducted a systematic review of peer-reviewed articles published between January 2022 and May 2025. We searched IEEE Xplore, ACM Digital Library, and SpringerLink, yielding 22,558 initial records. After carefully applying strict inclusion criteria, 125 papers were selected for the final analysis. We created a standardised data extraction form (i.e., using MS Excel) to gather bibliographic details, research emphasis, methodological strategies, datasets, evaluation criteria, and recognised constraints. We employed thematic analysis to develop a clear taxonomy. We identified five main research themes in our analysis: (1) ensemble and hybrid learning pipelines focused on performance optimisation (30 papers), (2) context-specific IDS designs for Internet of Things (IoT), cloud, and Software-Defined Networking (SDN) environments (34 papers), (3) data-centric engineering that deals with class imbalance and feature selection (20 papers), (4) deep neural architectures for representation learning (31 papers), and (5) trustworthiness concerns like adversarial robustness, zero-day detection, and Explainable AI (XAI) (10 papers). Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), and Random Forests are the most commonly used algorithms, often combined. Nonetheless, significant deficiencies remain: about 2% of papers incorporate XAI, only 4% focus on adversarial robustness, and none validate their models in real-world production settings. Denial-of-Service (DoS) and Distributed DoS (DDoS) attacks are the most common types in the literature, whereas Web attacks, ransomware, and advanced persistent threats remain poorly studied. The number of publications grows at an average of 30.2% annually, but the field still relies on legacy benchmark datasets rather than operational validation. Full article

Environmental contaminants (ECs) may impair skeletal health through endocrine disruption, oxidative stress, inflammation, and altered mineral homeostasis. This systematic review and meta-analysis evaluated the associations between EC exposure and osteoporosis-related outcomes in adults. PubMed/MEDLINE, Scopus, and Web of Science were searched in April 2026 for English-language studies published from 2016 onward. Study selection followed PRISMA guidelines, while methodological quality was evaluated using the Joanna Briggs Institute (JBI) checklist and Newcastle–Ottawa Scale (NOS). Evaluated contaminant classes encompassed ambient air pollution (PM2.5, PM10, PM1, NO₂, NO_x, SO₂, CO, O₃), per- and polyfluoroalkyl substances (PFAS), heavy metals and trace elements (e.g., cadmium, lead), pesticides, polycyclic aromatic hydrocarbons (PAHs), phthalates, volatile organic compounds (VOCs), persistent organic pollutants (POPs, dioxins, PCBs), brominated flame retardants (BFRs), and water disinfection by-products. Random-effects meta-analyses using restricted maximum likelihood estimation were conducted when at least 3 studies were available. Forty-eight studies involving 4,510,868 participants were included. Air contaminants showed the most consistent associations with reduced bone mineral density (BMD) and osteoporosis, followed by selected PFAS and cadmium exposures. Evidence for pesticides, PAHs, phthalates, volatile organic compounds VOCs, POPs, flame retardants, and water disinfection by-products was less consistent. PM2.5 exposure was associated with increased osteoporosis risk (OR = 1.24, 95% CI: 1.02–1.51), and to a lesser extend also PM10 (OR = 1.17, 95% CI: 0.96–1.42). The analyses showed extreme heterogeneity and sensitivity to influential studies. ECs, particularly fine particulate matter, may contribute to skeletal deterioration, although the findings require cautious interpretation. PROSPERO ID: CRD420261460078. Full article

Background: Platinum resistance remains a major therapeutic challenge in epithelial ovarian cancer (EOC). The tumour–stroma ratio (TSR) has emerged as a potential histopathological marker of tumour biology, but its clinical significance in different clinical settings remains unclear. We evaluated the association between stromal proportion, assessed using the TSR methodology, platinum resistance, and survival outcomes in patients undergoing primary debulking surgery (PDS) or neoadjuvant chemotherapy followed by interval debulking surgery (NACT+IDS). Methods: This retrospective study included 83 patients with epithelial ovarian cancer (EOC) who underwent either primary debulking surgery (PDS) or neoadjuvant chemotherapy followed by interval debulking surgery (NACT+IDS) between 2017 and 2024. Patients were analysed separately according to treatment strategy. Stromal proportion was assessed on primary surgical specimens in the PDS cohort and on pretreatment diagnostic biopsy specimens in the NACT+IDS cohort. For this retrospective analysis, platinum resistance was operationally defined as recurrence within six months after completion of first-line platinum-based chemotherapy. Survival outcomes were analysed using the Kaplan–Meier method, and univariable binary logistic regression was performed separately within the PDS and NACT+IDS cohorts to explore the association between stromal category and platinum resistance. Results: Platinum resistance occurred in 25.3% of patients. In the PDS cohort, stromal category was not associated with clinicopathological characteristics, platinum resistance, disease-free survival (DFS), or overall survival (OS). In the NACT+IDS cohort, patients in the stroma-high group had a numerically higher rate of platinum resistance than those in the stroma-low group (57.9% vs. 22.2%; Fisher’s exact p = 0.114). Based on univariable logistic regression analysis, the stroma-high group had higher estimated odds of platinum resistance (OR 4.81, 95% CI 0.78–29.40), although this finding did not reach statistical significance (p = 0.090). No significant association was observed between stromal category and survival outcomes in either cohort. Conclusions: In the NACT+IDS cohort, patients in the stroma-high group had a numerically higher rate of platinum resistance than those in the stroma-low group; however, this difference did not reach statistical significance. These results should be interpreted with caution due to the retrospective study design and limited sample size; however, they nonetheless support further investigation of pretreatment TSR as a simple and easily applicable histopathological parameter in EOC. Larger prospective multicentre studies are required to determine whether this exploratory signal is reproducible and clinically relevant. Full article

Cognitive impairment is a clinically important feature of schizophrenia that affects community functioning and responds poorly to pharmacological treatment. Cognitive remediation (CR) and physical exercise (PE) can each improve cognition, but whether their combination provides additive benefits remains uncertain. This systematic review examined cognitive outcomes of interventions combining CR and PE in adults with schizophrenia or other psychotic spectrum disorders. The protocol was registered in PROSPERO (ID: 639266). PubMed was searched for eligible studies, and findings were synthesized narratively because of heterogeneity in designs, interventions, comparators, and outcome measures. Risk of bias was assessed using RoB 2 for randomized trials and ROBINS-I for non-randomized studies. Eleven reports representing 10 underlying studies met the eligibility criteria. Combined CR and PE was associated with improvements in global cognitive performance and in domains including executive function, verbal memory, working memory, and processing speed. Some studies reported benefits over CR alone or PE alone, while three-arm trials provided the most direct, although not uniformly consistent, evidence of superiority over both monotherapies. Four randomized reports raised some concerns regarding risk of bias; three were judged to be at high risk, and all four non-randomized reports were judged to be at serious risk. The single-database search, methodological heterogeneity, and risk-of-bias concerns limit the certainty and generalizability of the findings. Combined CR and PE may provide additional cognitive benefits, but adequately powered trials with appropriate comparators and standardized outcomes are needed to establish additivity. Full article

Understanding how selected urban route conditions are associated with e-scooter riders’ visual attention and perceived workload can support human-factor evaluation of micromobility environments. This study examines visual attention and perceived workload during real-world e-scooter riding across three selected urban routes in Zagreb with different infrastructure and traffic-environment characteristics. A field-based experimental methodology was used, integrating eye-tracking with post-ride perceived workload assessment through the National Aeronautics and Space Administration Task Load Index (NASA-TLX) questionnaire. Twenty-eight adults, predominantly novice or occasional e-scooter riders, completed the three selected routes. Visual attention and perceived workload were examined across the selected routes; because only one route represented each route condition, the findings are interpreted as route-level evidence rather than general infrastructure-type effects. One-way repeated-measures analyses of variance (ANOVAs) showed that route condition had a significant effect on mean fixation duration and on the weighted NASA-TLX workload score, indicating that riders’ visual attention and perceived workload varied across the examined routes. To further examine these route-related differences, linear mixed-effects models were used as supporting analyses with participant ID as a random intercept. The results showed that R2 was associated with lower weighted workload than R1, while R3 was associated with higher weighted workload than R1. The route-complexity index was retained as an exploratory route-level descriptor rather than a validated continuous predictor. These findings highlight the importance of considering combined route, infrastructure, and traffic-environment characteristics when planning safer micromobility environments and developing appropriate regulations for e-scooter use. They also provide preliminary route-level human-factor evidence that can support future micromobility route evaluation and the development of human-centred guidance for infrastructure planning and e-scooter regulation. Full article

The design of healthcare facilities must comply with applicable regulatory requirements, which significantly influence functional, spatial, and performance aspects of healthcare buildings. However, compliance checking is still predominantly conducted through manual procedures, leading to time-consuming workflows and a high risk of errors. In this context, Computational Compliance Checking (CCC) offers the potential to enhance the efficiency and reliability of regulatory verification processes. The adoption of Building Information Modelling (BIM), supported by the open standard Industry Foundation Classes (IFC), enables the formalisation of Information Delivery Specifications (IDS) as machine-interpretable rules for model validation. Additionally, the buildingSMART Data Dictionary (bSDD) contributes to the semantic structuring and classification of information requirements. This study explores how the IDS standard can enhance CCC processes by supporting the automated verification of regulatory requirements. Complementary solutions based on Python are proposed to address cases where IDS alone proves insufficient. A methodology for the semi-automated compliance checking of healthcare regulatory requirements is developed and applied to a case study based on regional regulations in the Veneto Region (Italy). Tested on the inpatient ward of a hospital BIM model, the procedure processed all 35 accreditation requirements, 91.4% through IDS and 8.6% through complementary Python routines, revealing several design non-conformities. Full article

Background: Measurement reliability of the sagittal thoracic spine, e.g., thoracic kyphosis and balance, on radiographs has an unknown evidence base. This literature review aims to systematically identify and evaluate the reliability of biomechanical assessments of the sagittal thoracic spine on radiographs used in clinical practice. Methods: The study design was registered with PROSPERO (CRD42023431171). Chiropractic Biophysics Nonprofit, Inc (Eagle, ID, USA) funded this investigation. Inclusion criteria involved studies in English using human subjects, radiography, and reliability analysis of biomechanical analysis of the thoracic spine. Exclusion criteria involved animal and cadaveric studies, phantom mannequins, geometric studies, and non-radiographic studies. This review was conducted using the Peer Review of Electronic Search Strategies (PRESS) checklist to organize the search strategy. A combined approach using Medical Search Headings (MeSH) search terms and a systematic literature review (SLR) search strategy was used. This review followed the recommendations of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Databases searched included PubMed, CINAHL, Alt HealthWatch, Web of Science, and a grey literature search. Initial search results produced 990 results. A total of 658 records were screened by two independent reviews and 197 full text articles were assessed for eligibility. Results: Sixty-six studies were included in the final analysis and assessed for methodological quality and bias using the 11-item Quality Appraisal of Diagnostic Reliability (QAREL) tool. A total of 15 studies were of low methodological quality (high risk of bias), 28 were of moderate quality (moderate risk of bias), and 23 were of high quality (low risk of bias). Conclusions: This SLR found most articles investigating the reliability of biomechanical assessment of the sagittal thoracic spine on radiographs to be of moderate-to-high quality and show good-to-excellent reliability. Full article

Background: Episiotomy remains one of the most common obstetric procedures worldwide, and suboptimal healing of the perineal wound continues to cause pain, dehiscence, and long-term impairment of postpartum quality of life. Obstetric anal sphincter injury (OASI) carries the greatest risk of severe wound morbidity among perineal trauma types and is managed with prophylactic antibiotics and adjunctive measures; episiotomy, by contrast, is repaired without a comparable prophylactic package and is far more frequent, so that even a modest per-case complication rate carries a substantial absolute burden. Platelet-rich plasma (PRP), an autologous concentrate rich in growth factors, has demonstrated favourable effects on wound healing and scar quality across several surgical contexts, including caesarean section, raising the question of whether it could similarly benefit episiotomy repair. Methods: We conducted a structured, date-stamped search of PubMed, Scopus and Google Scholar, together with two trial registries (ClinicalTrials.gov, WHO ICTRP), forward and backward citation tracking, and reference-list screening of relevant reviews (final search 24 July 2026), to identify clinical studies of PRP for obstetric episiotomy or intrapartum perineal wound healing. Results: Five primary clinical studies and one systematic review were identified. One single-centre randomized controlled trial of 200 primiparous women, published in a regionally indexed journal covered by neither PubMed nor Scopus, evaluated PRP specifically for episiotomy wound healing and reported significantly lower REEDA, Vancouver and pain scores in the PRP arm; it was not prospectively registered, reported neither a sample-size calculation nor blinded outcome assessment, lost 12% of participants to follow-up, and contains internal inconsistencies in the reported data. The remaining evidence comprises one randomized trial of PRP for postpartum levator ani muscle recovery (a distinct target, with a null result), one non-randomized comparative study in grade III–IV intrapartum perineal laceration repair, two case reports, and one broad systematic review of PRP in pelvic floor disorders in which episiotomy was not analyzed as a distinct entity. Conclusions: PRP for episiotomy healing is therefore not wholly untested, but the single existing randomized trial is small, methodologically limited, not independently replicated, and effectively invisible to conventional database searching. We describe this evidence base and introduce PRP-EpiHeal (ClinicalTrials.gov ID: NCT07669285), designed to provide the first prospectively registered, adequately powered and assessor-blinded randomized evidence on this question, with REEDA scale assessment at six weeks postpartum as the primary outcome. Full article

Ensuring transparency and security in digital recruitment systems remains a critical challenge. This study proposes BC-XAIA, a unified framework that integrates blockchain, smart contracts, explainable artificial intelligence (XAI), and agile methodology to enable consistent, secure, and traceable recruitment decision-making. Smart contracts, implemented in Solidity and deployed using the Remix Ethereum IDE, automate key processes such as identity verification, data access control, and behavior monitoring, reducing reliance on centralized intermediaries. To support intelligent decision-making, multiple machine learning models, including Random Forest, Logistic Regression, and Support Vector Machine (SVM), were trained and evaluated on a recruitment dataset, with Random Forest achieving the highest performance, reaching an accuracy of 93%. To enhance transparency, SHAP and LIME were employed to provide both global and local interpretability of model predictions. Furthermore, agile methodology is embedded to drive continuous adaptation, iterative development, and stakeholder feedback throughout the recruitment lifecycle. Unlike existing recruitment systems that treat blockchain, AI, and explainability separately, BC-XAIA unifies these technologies within an agile and decentralized architecture. Overall, BC-XAIA establishes a secure, transparent, and explainable decentralized recruitment ecosystem that enhances trust, fairness, and intelligent decision-making in next-generation HR systems. Full article

Background/Objectives: Imaging-based markers of sarcopenia have emerged as potential indicators of frailty and predictors of clinical outcomes in patients with lower limb ischemia undergoing revascularization. However, the available evidence remains inconclusive. This systematic review aims to provide a comprehensive overview of current imaging-based approaches to sarcopenia assessment and to evaluate their relationship with clinical outcomes in this population. Methods: This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines and prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) database: ID CRD420251229527. A literature search was performed in PubMed and Scopus to identify studies published between 1 January 2020 and 1 August 2025. Studies assessing imaging-derived markers of sarcopenia in adults undergoing revascularization for lower limb ischemia were included. Due to heterogeneity in imaging methods, sarcopenia definitions and reported outcomes, a narrative synthesis was performed. Results: Nine studies involving 2563 patients were included in this review. Computed tomography (CT) was the predominant imaging modality used for sarcopenia assessment, with muscle measurements most frequently obtained at the level of the third lumbar vertebra (L3). Commonly reported imaging-derived markers included the skeletal muscle index, psoas muscle-based parameters and muscle radiodensity. Assessment strategies varied across studies, particularly with respect to measurement techniques, anatomical landmarks and the thresholds used to define sarcopenia. Most studies reported an association between imaging-derived low muscle mass and increased long-term mortality; however, independent associations after multivariable adjustment were less consistent. Findings regarding postoperative complications and limb-related outcomes were limited and inconsistent. Conclusions: Current evidence suggests that imaging-derived muscle markers may provide additional prognostic information regarding long-term mortality in patients undergoing revascularization for lower limb ischemia, while associations with postoperative complications and limb-related outcomes remain inconsistent. However, the small number of retrospective studies and substantial methodological heterogeneity limit their comparability, highlighting the need for standardization in future research. Full article

Building Information Modeling (BIM) has become a central paradigm in the architecture, engineering, and construction industry, enabling integrated management of design and construction data. Accessibility standards are essential to ensure inclusive and usable built environments, yet manual verification of ISO 21542:2021 criteria in BIM models is time-consuming and error-prone. Following the formal standardization of the Information Delivery Specification (IDS) in 2024, this study investigates its potential for systematic accessibility validation. Using the Design Science Research methodology, ISO 21542 door requirements were translated into IDS format and applied to the Urla Summer Villa case. The workflow integrated IDS Maker for XML rule definition, ArchiCAD for BIM, and BIMvision with the IDS Checker plugin for automated validation. The evaluation revealed an overall compliance rate of 50%, with non-compliances in door width, height, threshold, and handle parameters. Results indicate that IDS can detect accessibility inconsistencies early, provide transparent reporting, and strengthen quality assurance in BIM-based design processes. This study contributes to expanding the limited corpus of accessibility-focused IDS applications and highlights IDS as a practical tool for advancing inclusive design practices within digital construction workflows. The study is limited to a single case and door elements only, within a specific software ecosystem. Future research should extend IDS-based accessibility validation to ramps, stairs, elevators, and diverse BIM platforms to strengthen generalizability. Full article

Intrusion detection systems (IDSs) play a vital role in safeguarding modern computer networks against increasingly sophisticated and high-volume cyber threats. Recent progress in artificial intelligence, especially deep learning, has allowed IDSs to go from static rule-based systems to adaptive and data-driven security solutions. But traditional machine learning- and convolution-based IDSs often have trouble finding long-range dependencies and temporal correlations in large-scale network traffic. This makes detection less accurate and increases the number of false alarms. This challenge becomes more pronounced in heterogeneous and evolving network environments. To address this, experiments are conducted on two widely used benchmark datasets: CIC-IDS2017 for binary intrusion detection and CICIDS2018 for multiclass attack classification. These datasets represent realistic network traffic with diverse attack categories and severe class imbalance. The proposed methodology employs a Transformer-based intrusion detection framework incorporating sequence windowing, positional encoding, and multi-head self-attention to learn contextual traffic representations. The primary contribution of this study lies in systematically integrating sliding temporal windowing, positional encoding, and multi-head self-attention into flow-level intrusion modeling, accompanied by empirical ablation analysis and statistical validation across two large-scale CIC benchmark datasets. Performance is evaluated using accuracy, precision, recall, F1-score, ROC-AUC, and false alarm rate. Experimental results demonstrate that the proposed model achieves high detection accuracy, strong discriminative capability, and low false alarm rates across both datasets, confirming its effectiveness and scalability for next-generation cybersecurity applications. Full article