50716 research outputs found
Sort by
Enhanced Autoencoder Model for Robust Anomaly Detection in Financial Fraud with Imbalanced Data
No full textDetecting anomalies in financial fraud is challenging due to data imbalance and the variety of fraudulent techniques employed. Traditional autoencoders often exhibit bias towards the majority class, reducing their effectiveness in identifying anomalies. This paper introduces an enhanced autoencoder model that improves performance on imbalanced datasets by incorporating class-specific reconstruction losses and gradient clipping techniques. To avoid potential noise and bias, we did not employ the Synthetic Minority Over-sampling Technique. Instead, we evaluated our model using the Bank Account Fraud dataset and the BankSim dataset. Our findings reveal significant improvements in key performance metrics, including precision, recall, and F1-score, compared to traditional autoencoders. These results highlight the potential of advanced deep learning techniques to enhance anomaly detection systems.</p
Maintaining Assessment Validity in the GenAI Era:Insights from Human-Machine Interaction
No full textGenerative Artificial Intelligence (GenAI) has been altering the way that educational institutions and businesses operate. Leveraging the significant opportunities of GenAI and addressing its challenges are essential for institutions to maintain the competitiveness of their graduates and educational processes. For a few decades now, human-machine interaction scholars have been investigating how humans can properly utilise machine capabilities without over-reliance. Building on these insights, this paper proposes a framework for maintaining the effectiveness of higher education assessments amid the pervasive use of GenAI. The framework incorporates these insights into its five interconnected components: assessment design, establishing an understanding of purpose and process, effective monitoring, feedback, and reflection and planning for change. The framework allows for creating the environment and culture for the responsible use of GenAI while ensuring education quality and imposing academic integrity. We present two case studies to demonstrate the applicability of the proposed framework to diverse educational requirements. Our findings demonstrate how the framework facilitates fostering student creativity while maintaining confidence in the validity of the assessment system.</p
A systematic review of reviews on comprehensive community initiatives to prevent or reduce alcohol and other drug harms
No full textBackground and Aims: Comprehensive community initiatives (CCI) aimed at reducing or preventing alcohol or other drug (AOD) harms incorporate multiple initiatives delivered to whole communities to effect community-level change on sociocultural and environmental factors. CCIs have gained in popularity and have been subject to extensive research; however, CCIs comprise multiple initiatives and evidence for effectiveness by substance type has been mixed. This umbrella review aimed to synthesise information from published reviews to describe the combination of CCIs with the most consistent evidence for impact for each substance type. Method: We searched Embase/Medline, PsycINFO, Cochrane Database of Systematic Reviews, two online registries and hand searched references for English language reviews without date restriction and conducted an umbrella review using mixed methods synthesis (PROSPERO CRD42023432567). We considered all types of reviews focused on CCIs and addressing AOD use or harms. Two reviewers independently screened all articles and conducted full text review. Extraction of main results relating to CCI impact and quality assessment using AMSTAR-2 and SANRA was completed by two independent reviewers and corrected covered area analysis conducted. Results: We identified 87 reviews spanning three decades; 14 were rated high quality. Most reviews considered individual substances [alcohol (43 reviews) or tobacco (35 reviews) and rarely illicit drugs (16 reviews)], and some limited scope to ‘at-risk’ community members [young people (26 reviews) and First Nations (8 reviews)]. Although the evidence did not meet criteria for consistent impact, communities should consider implementing school-based (supported by 36 of 50 reviews) and parenting-related (29 of 37 reviews) activities that are supported by media campaign where feasible (29 of 51 reviews). CCIs have the most consistent impact on alcohol (supported by 24 of 32 reviews) and tobacco-related outcomes (22 of 35 reviews), though illicit drugs are yet to be adequately assessed. Conclusions: Although the available evidence regarding comprehensive community initiatives is largely inconsistent, the addition of parenting-related activities to existing school-based education campaigns is likely to improve effectiveness. A media campaign may extend their reach to those outside school settings. Future evaluation of CCIs should measure impact of activities in isolation where possible and incorporate process measures to gauge community engagement and empowerment.</p
The Impact of Different Types of Social Resources on Coping Self-Efficacy and Distress During Australia’s Black Summer Bushfires
No full textWhile social resources are known to promote positive psychological outcomes after disasters, little is known about the unique influence of different social resources on distress and coping during a disaster. This study examined the association between five social resources: sense of belonging, bushfire reciprocal support, emotional support, practical support and loneliness, and two psychological outcomes, distress and coping self-efficacy, during Australia’s 2019–2020 Black Summer bushfires. Survey data collected from 2611 bushfire-affected Australians in late 2020 was analysed using regression modelling. Higher perceived emotional and practical support and lower levels of loneliness predicted increased coping self-efficacy, and higher sense of belonging and lower loneliness predicted reduced distress. However, higher emotional and reciprocal support predicted higher distress after accounting for coping self-efficacy. The findings suggest having higher access to some social resources may not directly reduce distress but may reduce distress indirectly through increasing coping self-efficacy. While access to social resources, particularly bonding social capital, is likely important for supporting psychological response during disasters, the findings suggest this may be dependent on the perceived quantity, quality and expectations of these social resources. The findings indicate that different social resources interact with disaster-related psychological outcomes in distinct, complex and sometimes non-linear ways.</p
Integrating IMU Sensors and Dual-Task Timed Up and Go to Identify Biomarkers for Early Stage Parkinson’s Disease Detection
No full textParkinson’s disease (PD) is a progressive neurodegenerative disorder, where diagnosis is essential for effective management. However, current diagnostic approaches often lack sensitivity during the early clinical stages of the disease. This study investigates the potential of wearable inertial measurement units (IMUs) combined with machine learning (ML) techniques for detecting PD during mobility tasks. Data were collected from 28 individuals diagnosed with PD and 34 age-matched healthy controls while performing the timed up and go (TUG) test and its dual-task variants. A set of gait and mobility features was extracted and refined using two feature selection techniques. Four ML models, including support vector machine (SVM), logistic regression (LR), linear discriminant analysis (LDA), and gradient boosting (GB), were trained to classify participants. SVM and LR achieved the highest classification accuracy of 95%. Key discriminative features included turn duration, sit-to-stand time, and stand-to-sit time, which demonstrated potential as objective digital biomarkers for early stage PD detection. In addition, higher motor impairment scores were associated with increased shank range of motion and prolonged turn duration, particularly during dual-task conditions. These findings highlight the potential of wearable sensor-based mobility assessments combined with ML as a noninvasive, accessible, and reliable tool for early stage detection and monitoring of PD.</p
Podcasting comparative and international education: reflecting, reframing, or reorienting the field?
No full textPodcasts have increased in popularity in recent years as a form of knowledge production and dissemination, but scant research has examined how academic podcasts relate to their cognate fields. This study analysed in detail one academic podcast – FreshEd – to examine its content and how it may reflect, reframe, and/or reorient the field of comparative and international education (CIE). Data were collected on 294 podcast episodes, first analysed descriptively by mapping key information about the researchers featured on each, then thematically by inductively coding their focal content. The study is significant because it (1) examines how an academic podcast reflects or reframes a field (i.e. CIE); (2) explores the possibilities for academic podcasting to reorient the types of knowledge produced, and our relationships to them; and (3) exemplifies innovative methodological avenues for analysing new forms of knowledge dissemination beyond traditional scholarly boundaries.<br/
Prevalence of pelvic floor dysfunction in male athletes and its dose-dependency in high-intensity exercise:A scoping review
No full textBackground: This scoping review summarises literature on the prevalence and types of male pelvic floor dysfunction linked to increasing exercise intensity, with a secondary aim of exploring whether a dose-dependent relationship exists between the two. Methods: A comprehensive search of CINAHL, PEDro, PubMed, Scopus, and Web of Science databases was conducted up to January 2025 using MeSH terms and keywords relating to exercise and PFD. Results: The search yielded 6031 publications with 1199 duplicates removed prior to screening. After screening 4493 articles, 10 articles were eligible for inclusion. Substantial reporting differences of exercise-related variables were identified. Six studies reported prevalence rates of lower urinary tract symptoms (LUTS), including urinary incontinence, ranging from 3.8 % to 18.8 %. Only one study examined multiple pelvic floor issues, identifying anorectal incontinence in 61.8 % of participants. Studies investigating erectile dysfunction (ED) and chronic prostatitis/pelvic pain (CP/CPPS) reported hazard and odds ratios. When reporting as ratios for ED, CP/CPPS, studies identified that higher weekly MET-minutes of physical activity were slightly more protective than sedentary or lower physical activity levels. The cross-sectional studies reporting LUTS as a sample percentage identified a dose-dependent relationship between weekly MET-minutes and the prevalence of male LUTS, based on both minimum and maximum weighted averages (R = 0.86 vs. R = 0.87). Conclusions: Athletic males may experience higher rates of LUTS than the general population. A dose-dependent relationship between MET-minutes per week and male LUTS indicates that men who exercise intensely at higher volumes may be at greater risk of symptomology, whereas risk of ED and CP/CPPS may be reduced.</p
Approaches to community sport policy analysis: a scoping review of key tenets of the Advocacy Coalition Framework
No full textCommunity sport policy is characterised as complex, due in part to the competing interests of the stakeholders involved and the marginalisation of community sport clubs (CSCs). There is too great a focus on policy creation rather than effective policy analysis, and few studies have assessed the application of mainstream meso-level theories in sport policy studies. Further, there are weaknesses in policy processes that apply to community sport and an absence of a robust framework that factors in the viewpoints of all stakeholders. This study seeks to outline the potential for the Advocacy Coalition Framework (ACF) to fill that void. Through a scoping review, this paper identifies the application of key characteristics of the ACF in policy analysis relating to sport with a view to greater consideration of the framework in policy processes going forward. A database search of articles published from 2004 to 2025 identified the application of fundamental ACF constructs within sport policy research. A range of studies (n = 114) connected with sport policy contained relevant contextual determinants, suggesting that ACF principles are increasingly applicable in sport policy analysis. Overall, this paper provides the basis for the appraisal of community sport policy that demonstrates the ACF’s potential, through advocacy, for a more thorough and relevant approach to policy in which CSCs may become less sidelined in the process.</p
Impact of an Overload Period on Heart Rate Variability, Sleep Quality, Motivation, and Performance in High-level Swimmers:Use of Explainable Artificial Intelligence (XAI) to Assess Training Load Variations
No full textBACKGROUND: Understanding the impact of training sessions on physiological, psychological, and immunological responses is crucial for adequate training periodization and preventing negative influences on health, training, and performance.OBJECTIVES: To characterize the responses of heart rate variability (HRV), sleep time and quality, motivation, dry-land strength, and swimming performance to an overload period of three consecutive 7-day cycles (cycles 1, 2, and 3) with different training intensity and volume dynamics. Secondly, to test the capability of HRV to assess daily variation in training loads on the basis of explainable artificial intelligence (XAI) models.METHODS: A total of 14 high-level swimmers (4 males and 10 females, aged 17.5 ± 1.5 years) were monitored via an orthostatic test, Hooper index, sleep questionnaires, and rating of perceived exertion (RPE) of each training session. The self-reported and prescribed training loads were compared. At the beginning of each cycle and at the end of cycle 3, swimmers completed anthropometric testing, countermovement jumps, hand-grip strength tests, and a 5 × 200 m incremental protocol.RESULTS: High-level swimmers accurately perceived their daily training loads. However, differences between the training and RPE loads emerged on weekends, indicating that physiological and psychological loads have different influences and should be considered simultaneously when characterizing training loads. The overload period was characterized by an increase in both training (27%) and RPE (20%) loads without eliciting a negative effect on sleep quantity and quality. During the overload period, supine (F 2.18 = 3.448, η 2 = 0.28; p = 0.05) and standing (F 2.18 = 3.809, η 2 = 0.30; p = 0.04) mean heart rate (HR) increased and supine log root mean square of the successive differences (LnRMSSD; F 2.18 = 4.379, η 2 = 0.33; p = 0.028) and maximal blood lactate (F 3.27 = 3.441, η 2 = 0.28; p = 0.03) decreased during and after cycle 3 (respectively). Dry-land and swimming performances were maintained, indicating that the autonomic nervous system appears to be more sensitive (XAI models r 2 = 0.91 and 0.9) to changes in acute/short-term training load. CONCLUSIONS: HRV indices, particularly supine RMSSD and mean HR, were the most sensitive markers of training load variation, while sleep, strength and power, and swimming performance remained stable. HRV can be employed as a practical tool for monitoring training responses and managing training loads in competitive swimmers.</p
Finite Element Modeling of Residual Stress Formation during Nanosecond Laser Ablation of Stainless Steel
No full textLaser cleaning has recently gained attention as a non-contact, precise, and environmentally friendly method for removing contaminants, oxides, or coatings from surfaces using controlled laser ablation. However, even at low fluence, where no visible damage is observed, thermal effects from nanosecond laser pulses can induce residual stress, potentially affecting the long-term reliability of the material. This study introduces a finite element method model to predict residual stress formation in single-pulse laser ablation of stainless steel, providing detailed insight into the localized thermomechanical response. The model captures individual pulse effects, including stress evolution, thermal expansion, and localized plastic deformation. Experimental validation via X-ray diffraction is also performed. Based on this model, a parametric study is conducted to investigate the influence of pulse duration and fluence on residual stress development. The results reveal how thermal input parameters affect stress field formation and redistribution, even in the absence of visible material removal. These findings enhance understanding of stress generation in laser-processed metals and offer a framework for predicting residual stress and identifying truly damage-free thresholds in metallic substrates.</p