8333 research outputs found
Sort by
Predictors of Quality of Life among Single Mothers and Single Fathers
No full textObjectives: A growing number of studies suggest that loneliness and self-stigma are important predictors of well-being and quality of life (QOL) among single parents. However, much of this research has focused primarily on single mothers and less is understood about how these factors impact single fathers. The overall aim of this research is to examine if loneliness and self-stigma predict QOL among single parents and identify if there are gender differences.
Method: A quantitative approach using an anonymous online questionnaire through Microsoft forms recruited and analysed 93 single parents (70 women and 23 men) examining the loneliness, self-stigma and QOL.
Results: As feelings of loneliness increase, QOL decreases. Furthermore, gender differences are evident with single parent men reporting lower loneliness and self-stigma scores when compared to single mothers (p<.001). Additionally, the QOL levels among single fathers were significantly higher than single mothers indicating that single mothers had higher feelings of loneliness and self-stigma resulting in lower QOL levels.
Conclusion: The present study supports previous findings and extends these to include gender differences in loneliness, self-stigma, and QOL. Future research should examine these variables in larger, population representative samples, and include qualitative research to further examine individual beliefs and experiences surrounding single parenthood
EP-MPCHS: Edge Server-Based Cloudlet Offloading Using Multi-Core and Parallel Heap Structures
No full textThe increasing computational demands of mobile applications, like image caption generators and Google Lens, result in higher memory and Random Access Memory (RAM) usage. Thus, to offboard the computational workload of these applications, cloud-based edge server frameworks have been in demand lately. In the age of cloud-based computing, the study aims to create a hybridized cloudlet placement algorithm that caters to reducing latency, increasing bandwidth, reducing network flow pressure, and optimizing the edge server resources. The primary condition of the utilized placement algorithm is to prioritize the cloudlets enabling the reduction of latency, increase of bandwidth, and CPU utilization. This study proposes Edge Priority Placement using Multi-Core and Parallel Heap Structures (EP-MCPHS) utilizing the min heap prioritization technique to deduce the placement of each cloudlet. This incorporates the priority queue-based resource allocation system which ensures that the optimal process is selected to the minimum available edge server allowing the reduction of resource utilization, increased latency, decreased network flow, and enhanced bandwidth. The algorithm also reinforces the technique with a multi-latent parallel head provisioning or Ph.C. with parallel processing, allowing reduced process starvation for the non-processing cloudlets. EP-MCPHS reduces the end time for cloudlet processing by 27.36%, increases bandwidth by 71.27%, and data flow by 24.81%. This study incorporates the SimPy framework for simulation testing using the EdgeSimPy framework for edge server simulation. The study compares the results achieved by the architecture with the Min-Max fairness algorithm. It provides statistical testing across time intervals to showcase the ability of the placement algorithm even with increasing workload
Exploring Self-Esteem, Well-Being, Cannabis and Cocaine Involvement Among Adults in Ireland
No full textAims: This study aims to explore the interplay between self-esteem (SE), well-being (WB), cannabis and cocaine involvement amongst adults in Ireland in the context of the normalisation of drug use, self-medication and the theory of planned behaviour (TPB).
Method: A survey was administered to participants (n=185) through Microsoft Forms assessing psychological factors (SE and WB), cannabis and cocaine use patterns. Variables were assessed using the cross-sectional survey of 185 participants using standardised measures (Rosenberg Self-Esteem Scale, WHO-5 Well-Being Index, and WHO Alcohol, Smoking and Substance Involvement Screening Test).
Results: Age was positively correlated with both cannabis and cocaine involvement. No significant relationships were found between substance use (SU) and socioeconomic factors (education, occupation). Gender differences were found in males having higher cannabis use than females, where cocaine involvement had not significant differences. SE showed a negative correlation with cannabis and cocaine use; notably once age was controlled, neither SE nor WB predicted SU.
Conclusion: These finding fits with the normalisation thesis that suggests shifts in cultural acceptance regarding what is referred to as softer, less dangerous substances, such as cannabis. In contrast to traditional stereotypes, gender differences only exist for cannabis involvement in this group. The finding supports the SE and WB enhancement over abstinence-based policies in the context of a youth-oriented intervention. This study encourages more holistic strategies to tackle SU in Ireland’s changing cultural milieu, by taking on board psychological vulnerabilities and the way society operates
Dating Application Fraud Profile Detection and Analysis using Data Mining
No full textWith the popularity that online dating application has gained, safety of users has been compromised by the increase of fraudulent activities such as catfishing, identity theft, and financial scams. This work focuses on detecting and analysing fraudulent profiles using data mining and machine learning techniques. The K-Means clustering method of unsupervised learning techniques is used in this study to detect anomalies in user profiles. Text-mining approaches like sentiment analysis and topic modelling are employed to identify the deception patterns in the detected anomalies within the profile. In addition, Support Vector Machine (SVM) as classification models are used to forecast fraudulent profiles. Experimental results show that combining clustering, sentiment analysis, and classification is more accurate at detecting fraud resulting in higher precision in SVM. Future aspects will aim to improve the class-balance and integrate advanced NLP models including real-time datasets can make fraud detection in dating applications reliable
Mind over Myths: Exploring the Public Endorsement of Mental Illness Misconceptions and their Associated Factors
No full textObjectives: Research on the formation and persistence of misconceptions, particularly in relation to mental illness, highlights the need to explore the broader network of beliefs that collectively influence rational thinking. The present study investigated whether pseudoscientific beliefs, paranormal thinking and cognitive reflection each predict misconceptions about mental illness, and examined their endorsement rate among the general public.
Method: A total of 157 participants completed an online questionnaire examining their endorsement of mental illness misconceptions, as well as measures assessing their pseudoscientific beliefs, paranormal thinking and cognitive reflection. Preliminary analyses were carried out to account for the influence of several covariates (age, education, history of mental health diagnosis, religious and political affiliation), following a hierarchical multiple regression analysis using SPSS version 28.0.1.1.
Results: The findings of the present study support the hypothesis that higher levels of pseudoscientific thinking are positively correlated with increased misconceptions about mental illness and its treatment. Contrary to expectations, neither paranormal beliefs nor cognitive reflection significantly predicted the endorsement of mental illness misconceptions, thus rejecting these hypothesized associations. However, lower education levels emerged as a significant contributing factor.
Conclusion: These results emphasize the importance of addressing pseudoscientific beliefs in an effort to reduce misconceptions about mental illness. Improving the ability to differentiate between pseudoscience and evidence-based science, especially in earlier stages of education, may help prevent the formation of flawed inferential frameworks, thereby reducing susceptibility to misconceptions endorsement. This approach may subsequently contribute to a reduction in stigmatization and negative societal attitudes toward mental illness
Investigating the Relationship Between Loneliness and Mental Health Help-Seeking Behaviour
No full textAims: The present study sought to explore the relationship between loneliness and help-seeking within the general population and hypothesised a significant correlation. This study compared differences of help-seeking between men and women and hypothesised that women are more likely to seek help. Lastly this study also investigated how age, gender and perceived social support predicted help-seeking and hypothesised that these would be significant predictors of help-seeking. Research has shown the profound harmful effects that loneliness can have on overall health. Therefore, the present study sought to analyse the role that help-seeking can have to mitigate loneliness.
Method: A survey was administered to participants (n=103), and they were recruited online via social media, and messaging platforms. This survey consisted of demographic questions of age and gender, the Revised UCLA Loneliness scale was used to analyse loneliness and social isolation, The General Help-Seeking Questionnaire (GHSQ) was used as a self-report measure to investigate future help seeking intentions. The Multidimensional Survey of Perceived Social Support (MSPSS) was used to examine levels of support from family, friends and significant others.
Results: Findings did not identify a statistically significant correlation between loneliness and help seeking. Follow up independent t-tests found that women show significantly higher levels of help-seeking compared to men. Findings from the multiple regression analysis found that the model explained 29.9% of variance in help seeking and that perceived social support, but not age or gender, was significantly predictive of help-seeking.
Conclusion: This study challenges the idea that lonely individuals will naturally seek help and suggests clinical implications aimed at more proactive outreach rather than waiting for individuals to seek-help
A Federated Blockchain Security for MEC-enabled IoT Networks in Industrial 5.0
No full textSecure communication is vital for the Industrial Internet of Things (IIoT) in the fifth revolution. However, resource constraints of most IIoT devices prevent the use of conventional security models. Multi-access mobile edge computing (MEC) offers a solution by bringing computational resources closer to the network edge, but centralized security solutions limit scalability and flexibility in MEC-enabled IIoT systems. Additionally, the distributed nature of MEC servers poses data privacy and security challenges. This paper introduces a novel network security approach for IIoT using federated blockchain (FB) and a machine learning-based verification system (ML). MEC optimizes the FB model, ensuring data integrity and confidentiality between the IIoT's local network cluster and external devices. Public key cryptography secures data shared within the local network cluster, and the ML-based verification model validates IIoT devices requesting encryption key-pair updates and joining the MEC's FB. This integrated approach surpasses conventional security solutions by enhancing scalability, data privacy, and adaptability to the evolving IIoT network topology. The methodology is implemented and evaluated using a realistic IoT testbed, demonstrating improved network security while maintaining MEC-enabled IIoT system performance and scalability
Integrating Data Mining, Statistics, and Machine Learning for Enhanced Credit Risk Scoring
No full textThe banking and financial services sector has transformed its credit risk assessment process through the fast-moving development of data analytics with machine learning and AI applications to determine borrowing capacity and set lending restrictions. Numerous machine learning algorithms such as XGBoost and CatBoost and HistGradientBoosting supplant traditional assessment tools because they deliver precise credit risk evaluations together with flexible adaptation and detailed analytics. This research examines credit risk assessment challenges that encompass predictive accuracy together with fairness requirements and regulatory standards. This research applies advanced algorithms together with explainable AI (XAI) methods SHAP and LIME to enhance model interpretation capabilities while establishing trust between stakeholders. The framework incorporates advanced predictive models in a single system which delivers fairness alongside ethical features to satisfy developing regulatory requirements and social norms. Through feature engineering combined with two-stage bias mitigation strategies applied during data preprocessing and model construction the research demonstrates pathways toward demographic group inclusivity. The framework shows practical use in financial reality through its ability to process data in real-time for largescale datasets. The proposed approach delivers enhanced predictive accuracy alongside transparency and fairness that allows financial institutions to maintain detailed social equity decision-making and accountably through scientific rigors which properly integrate technology with ethical and societal standards
EDGE360: Edge-Enabled Multi-Agent DRL for Region-Aware Rate Adaptation Solution to Enhance Quality of 360° Video Streaming
No full textOptimal tile-based bitrate allocation improves the Quality of Experience (QoE) for adaptive 360° video streaming across multiple clients in heterogeneous network environments; however, it is challenging as it implies accurate viewport prediction, finest tile-based bitrate reservation, and maintaining QoE fairness, particularly under constrained network conditions. This paper proposes a strategy named EDGE360, that employs an edge-driven Multi-Agent Deep Reinforcement Learning (MADRL) solution for rate adaptation to improve the joint QoE in DASH-based rich media content delivery based on adaptive viewport prediction and Video Multi-method Assessment Fusion (VMAF) corresponding tiling granularity selection. Cooperative strategies among agents in the central critic network are crucial for addressing the complexity of network instances at the edge and optimizing media streaming bitrate assignment in multiple-client scenarios. Therefore, EDGE360 aims to implement the Counterfactual Multi-Agent Policy Gradients (COMA) based on 5G network traces to train agents in policies that optimize individual client QoE and fairness among clients, resulting in an improved rich streaming experience. At the edge, a tile-based quality monitor evaluates viewport trajectories, buffer status, and network throughput, employing deep learning to forecast optimal tile bitrate allocation, which is formulated as an MDP and solved with MADRL. Based on extensive experimentation, EDGE360 surpasses state-of-the-art adaptive bitrate algorithms by achieving the highest average reward, outperforming RAPT360, 360SRL, and BOLA360 by 8.12%, 11.86%, and 18.00%, respectively, demonstrating superior convergence and refinement
Senior Coding Framework for Enhancing Python Skills in Students Aged 13-16
No full textThe Senior Coding Club (SCC) was designed to teach Python programming to students aged 13 to 16, in the National College of Ireland, Dublin, Ireland, aiming to enhance their skills and confidence in STEM fields. Raspberry Pis were used as the platform for teaching. The comprehensive framework included a pre-programme phase to assess readiness, a structured delivery phase with a balanced focus on theoretical and practical learning, and a postprogramme phase for impact analysis. Core activities included hands-on coding labs, daily assessments, and a culminating hackathon to showcase student projects, with active parental involvement throughout. For students, the program led to significant gains in confidence and interest in STEM subjects. More inclusive attitudes toward STEM was observed. Parents reported enhanced confidence in supporting their children's STEM education. They became more aware of local STEM opportunities and displayed greater positivity about their children's abilities and interests. The SCC boosted students' confidence, with 82 % feeling “Quite confident” or “Very confident” in computer skills (up from 29%) and 71 % in coding (up from almost none). Among parents, 100 % agreed it improved STEM skills, and 78.95 % noted increased interest in STEM, underscoring SCC's broad success