Search Results (649)

The lack of fault data is still a major concern in the area of smart maintenance, as these data are required to perform an adequate diagnostics and prognostics of the system. In some instances, fault data are adequately collected, even though the fault labels are missing. Accordingly, the development of methodologies that generate these missing fault labels is required. In this study, Markov-CVAELabeller is introduced in an attempt to address the lack of fault label challenge. Markov-CVAELabeller comprises three main phases: (1) image encoding through the application of the first-order Markov chain, (2) latent space representation through the consideration of a convolutional variational autoencoder (CVAE), and (3) clustering analysis through the implementation of k-means. Additionally, to evaluate the accuracy of the method, a convolutional neural network (CNN) is considered as part of the fault classification task. A case study is also presented to highlight the performance of the method. Specifically, a hydraulic test rig is considered to assess its condition as part of the fault diagnosis framework. Results indicate the promising applications that this type of methods can facilitate, as the average accuracy presented in this study was 97%. Full article

This study introduces a robust offline system for 2D indoor navigation, developed to address common challenges such as complex layouts and connectivity constraints in diverse environments. The system leverages advanced spatial modeling techniques to optimize pathfinding and resource efficiency. Utilizing a structured development process, the proposed solution integrates lightweight data structures and modular components to minimize computational load and enhance scalability. Experimental validation involved a comparative approach: traditional navigation methods were assessed against the proposed system, focusing on usability, search efficiency, and user satisfaction. The results demonstrate that the offline system significantly improves navigation performance and user experience, particularly in environments with limited connectivity. By providing intuitive navigation tools and seamless offline operation, the system enhances accessibility for users in educational and other complex settings. Future work aims to extend this approach to incorporate additional features, such as dynamic adaptability and expanded application in sectors like healthcare and public services. Full article

Background/Objectives: With advancements in Large Language Models (LLMs), counseling chatbots are becoming essential tools for delivering scalable and accessible mental health support. Traditional evaluation scales, however, fail to adequately capture the sophisticated capabilities of these systems, such as personalized interactions, empathetic responses, and memory retention. This study aims to design a robust and comprehensive evaluation scale, the Comprehensive Evaluation Scale for LLM-Powered Counseling Chatbots (CES-LCC), using the eDelphi method to address this gap. Methods: A panel of 16 experts in psychology, artificial intelligence, human-computer interaction, and digital therapeutics participated in two iterative eDelphi rounds. The process focused on refining dimensions and items based on qualitative and quantitative feedback. Initial validation, conducted after assembling the final version of the scale, involved 49 participants using the CES-LCC to evaluate an LLM-powered chatbot delivering Self-Help Plus (SH+), an Acceptance and Commitment Therapy-based intervention for stress management. Results: The final version of the CES-LCC features 27 items grouped into nine dimensions: Understanding Requests, Providing Helpful Information, Clarity and Relevance of Responses, Language Quality, Trust, Emotional Support, Guidance and Direction, Memory, and Overall Satisfaction. Initial real-world validation revealed high internal consistency (Cronbach’s alpha = 0.94), although minor adjustments are required for specific dimensions, such as Clarity and Relevance of Responses. Conclusions: The CES-LCC fills a critical gap in the evaluation of LLM-powered counseling chatbots, offering a standardized tool for assessing their multifaceted capabilities. While preliminary results are promising, further research is needed to validate the scale across diverse populations and settings. Full article

Survival analysis is a crucial statistical technique used to estimate the anticipated duration until a specific event occurs. However, current methods often involve discretizing the time scale and struggle with managing absent features within the data. This becomes especially pertinent since events can transpire at any given point, rendering event analysis a continuous concern. Additionally, the presence of missing attributes within tabular data is widespread. By leveraging recent developments of Transformer and Self-Supervised Learning (SSL), we introduce SSL-SurvFormer. This entails a continuously monotonic Transformer network, empowered by SSL pre-training, that is designed to address the challenges presented by continuous events and absent features in survival prediction. Our proposed continuously monotonic Transformer model facilitates accurate estimation of survival probabilities, thereby bypassing the need for temporal discretization. Additionally, our SSL pre-training strategy incorporates data transformation to adeptly manage missing information. The SSL pre-training encompasses two tasks: mask prediction, which identifies positions of absent features, and reconstruction, which endeavors to recover absent elements based on observed ones. Our empirical evaluations conducted across a variety of datasets, including FLCHAIN, METABRIC, and SUPPORT, consistently highlight the superior performance of SSL-SurvFormer in comparison to existing methods. Additionally, SSL-SurvFormer demonstrates effectiveness in handling missing values, a critical aspect often encountered in real-world datasets. Full article

Massive Open Online Courses (MOOCs) have become increasingly prevalent in higher education, with the COVID-19 pandemic further accelerating their integration, particularly in developing countries. While MOOCs offered a vital solution for educational continuity during the pandemic, factors influencing students’ sustained engagement with them remain understudied. This longitudinal study examines the factors influencing learners’ sustained engagement with ThaiMOOC, incorporating demographic characteristics, usage log data, and key predictors of adoption and completion. Our research collected primary data from 841 university students who enrolled in ThaiMOOC as a mandatory curriculum component, using online surveys with open-ended questions and post-course usage log analysis. Logistic regression analysis indicates that adoption intention, course content, and perceived effectiveness significantly predict students’ Actual Continued Usage (ACU). Moreover, gender, prior MOOC experience, and specific usage behaviors emerge as influential factors. Content analysis highlights the importance of local language support and the desire for safety during the COVID-19 pandemic. Key elements driving ACU include video design, course content, assessment, and learner-to-learner interaction. Full article

This study compared the efficacy of AI neuroscience tools versus traditional design methods in enhancing viewer engagement with political campaign materials from the Harris–Trump presidential campaigns. Utilising a mixed-methods approach, we integrated quantitative analysis employing AI’s eye-tracking consumer behaviour metrics (Predict, trained on 180,000 screenings) with an AI-LLM neuroscience-based marketing assistant (CoPilot), with 67,429 areas of interest (AOIs). The original flyer, from an Al Jazeera article, served as the baseline. Professional graphic designers created three redesigned versions, and one was done using recommendations from CoPilot. Metrics including total attention, engagement, start attention, end attention, and percentage seen were evaluated across 13–14 areas of interest (AOIs) for each design. Results indicated that human-enhanced Design 1 with AI eye-tracking achieved superior overall performance across multiple metrics. While the AI-enhanced Design 3 demonstrated strengths in optimising specific AOIs, it did not consistently outperform human-touched designs, particularly in text-heavy areas. The study underscores the complex interplay between neuroscience AI algorithms and human-centred design in political campaign branding, offering valuable insights for future research in neuromarketing and design communication strategies. Python, Pandas, Matplotlib, Seaborn, Spearman correlation, and the Kruskal–Wallis H-test were employed for data analysis and visualisation. Full article

Clinical supervision is essential for trainees, preventing burnout and ensuring the effectiveness of their interventions. AI technologies offer increasing possibilities for developing clinical practices, with supervision being particularly suited for automation. The aim of this study is to evaluate the feasibility of using ChatGPT-4 as a supervisory tool in psychotherapy training. To achieve this, a clinical case was presented to three distinct groups (untrained AI, pre-trained AI, and qualified human supervisor), and their feedback was evaluated by Gestalt psychotherapy trainees using a Likert scale rating of satisfaction. Statistical analysis, using the statistical package SPSS version 25 and applying principal component analysis (PCA) and one-way analysis of variance (ANOVA), demonstrated significant differences in favor of pre-trained AI feedback. PCA highlighted four components of the questionnaire: relational and emotional (C1), didactic and technical quality (C2), treatment support and development (C3), and professional orientation and adaptability (C4). The ratings of satisfaction obtained from the three kinds of supervisory feedback were compared using ANOVA. The feedback generated by the pre-trained AI (f2) was rated significantly higher than the other two (untrained AI feedback (f1) and human feedback (f3)) in C4; in C1, the superiority of f2 over f1 but not over f3 appears significant. These results suggest that AI, when appropriately calibrated, may be an appreciable tool for complementing the effectiveness of clinical supervision, offering an innovative blended supervision methodology, in particular in the area of career guidance. Full article

Mental health illness is the single biggest cause of inability within the UK, contributing up to 22.8% of the whole burden compared to 15.9% for cancer and 16.2% for cardiovascular disease. The more extensive financial costs of mental ailments in Britain have been evaluated at British Pound Sterling (GBP) 105.2 billion each year. This burden could be decreased with productive forms and utilization of computerized innovations. Electronical health records (EHRs), for instance, could offer an extraordinary opportunity for research and provide improved and optimized care. Consequently, this technological advance would unburden the mental health system and help provide optimized and efficient care to the patients. Using natural language processing methods to explore unstructured EHR text data from mental health services in the National Health Service (NHS) UK brings opportunities and technical challenges in the use of such data and possible solutions. This descriptive study compared technical methods and approaches to leverage large-scale text data in EHRs of mental health service providers in the NHS. We conclude that the method used is suitable for mental health services. However, broader studies including other hospital sites are still needed to validate the method. Full article

Virtual museums powered by virtual reality (VR) technology serve as innovative platforms for cultural preservation and education, combining accessibility with immersive user experiences. While gamification has been widely explored in educational and entertainment contexts, its impact on user experiences in virtual cultural heritage museums remains underexplored. Prior research has focused primarily on engagement and enjoyment in gamified virtual environments but has not sufficiently distinguished between hedonic (pleasure-driven) and eudaimonic (meaning-driven) experiences or their impact on learning outcomes. This study aims to address this gap by comparing gamified and non-gamified virtual museum designs to evaluate their effects on hedonic and eudaimonic experiences, knowledge acquisition, and behavioral engagement. Using a quasi-experimental approach with 70 participants, the findings indicate that gamification significantly enhances hedonic experiences, including enjoyment, engagement, and satisfaction, while fostering prolonged interaction and deeper exploration. However, eudaimonic outcomes such as personal growth and reflection did not exhibit statistically significant differences. These results underscore the potential of gamified VR environments to balance entertainment and educational value, offering insights into user-centered design strategies for virtual museum systems that bridge technology, culture, and engagement. Full article

Despite the growing presence of digital Virtual YouTubers (Vtubers) in educational settings, there is limited empirical evidence on their effectiveness in language acquisition. In this investigation, we delved into the realm of digital education to assess how the visual fidelity of digital Vtuber avatars affects the acquisition of Mandarin Chinese characters by beginners. Through incorporating a diverse array of digital Vtubers, ranging from simple two-dimensional figures to complex three-dimensional models, we explored the relationship between digital Vtuber design and learner engagement and efficacy. This study employed a randomized tutorial distribution, immediate post-tutorial quizzing, and a realism scoring rubric, with statistical analysis conducted through Pearson correlation. The analysis, involving 608 participants, illuminated a clear positive correlation: digital Vtubers with higher levels of realism significantly enhanced learning outcomes, underscoring the importance of visual fidelity in educational content. This research substantiates the educational utility of digital Vtubers and underscores their potential in creating more immersive and effective digital learning environments. The findings advocate for leveraging sophisticated digital Vtubers to foster deeper learner engagement, improve educational achievement, and promote sustainable educational practices, offering insights for the future development of digital learning strategies. Full article

Fault diagnostics in internal combustion engines (ICEs) is vital for optimal operation and avoiding costly breakdowns. This paper reviews methodologies for ICE fault detection, including model-based and data-driven approaches. The former uses physical models of engine components to diagnose defects, while the latter employs statistical analysis of sensor data to identify patterns indicating faults. Various methods for ICE fault identification, such as vibration analysis, thermography, acoustic analysis, and optical approaches, are reviewed. This paper also explores the latest approaches for detecting ICE faults. It highlights the challenges in the diagnostic process and ways to enhance result accuracy and reliability. This paper concludes with a review of the progress in fault identification in ICE components and prospects, highlighted by an experimental investigation using 16 machine learning algorithms with seven feature selection techniques under three load conditions to detect faults in a four-cylinder ICE. Additionally, this study incorporates advanced deep learning techniques, including a deep neural network (DNN), a one-dimensional convolutional neural network (1D-CNN), Transformer and a hybrid Transformer and DNN model which demonstrate superior performance in fault detection compared to traditional machine learning methods. Full article

Online fraud threatens the well-being of older adults, with disparities in digital literacy and socioeconomic conditions amplifying their vulnerability. This study examined digital literacy and fraud victimization behavior among older adults in urban and rural settings, identifying key factors influencing victimization and its consequences. This cross-sectional analytical study, using multi-stage sampling, included 864 participants from Southern Thailand. The findings revealed that 46.3% of participants had adequate digital literacy, while 75.3% experienced fraud victimization, with higher rates of health impacts in rural areas. Higher age (Adjusted Odds Ratios; AOR: 1.83, p = 0.004), income (AOR: 2.28, p = 0.003), and rural residence (AOR: 3.03, p < 0.001) were significantly associated with an increased likelihood of fraudulent victimization. Conversely, being non-Buddhist (AOR: 0.47, p = 0.001) and having an adequate digital literacy (AOR: 0.50, p < 0.001) were protective factors. Fraud victimization significantly affected older adults’ health, with 29.5% reporting the following adverse outcomes: physical (AOR: 5.55), emotional (AOR: 7.80), social (AOR: 4.97), and overall heightened health risks (AOR: 7.71, p < 0.001). This research highlights the importance of improving digital literacy, fostering community awareness, and implementing tailored fraud-prevention strategies to protect older adults. This study provides a foundation for evidence-based policies aimed at mitigating digital risks and enhancing older adults’ well-being in the digital era. Full article

The fourth industrial revolution has ushered in a new era in which technology is seamlessly integrated into daily life. The digital transformation has created new media formats that require the development of robust digital skills to navigate this landscape. By utilising the Youth Digital Skills Indicator (yDSI) and integrating it with the Digital Competence Framework for Citizens (DigComp 2.2), this research examines media habits and digital competences among Croatian youth aged 10–24, corresponding to Generations Alpha and Z. A sample of 231 participants across three competence domains—information literacy, security and communication—revealed statistically significant generational differences in the first two areas of digital skills. Furthermore, gender-based analyses, conducted using the Mann–Whitney U-test and Spearman correlations for Likert scale responses, showed no significant differences. These findings deepen our understanding of digital natives, how media habits evolve and influence their digital skills, highlighting the need for more tailored strategies to enhance their competences and bridge generational gaps. Full article

Planning in mass-customization supply and manufacturing processes is a complex process that requires continuous planning and optimization to minimize time and cost across a wide variety of choices in large production volumes. While soft computing techniques are widely used for optimizing mass-customization products, they face scalability issues when handling large datasets and rely heavily on manually defined rules, which are prone to errors. In contrast, machine learning techniques offer an opportunity to overcome these challenges by automating rule generation and improving scalability. However, their full potential has yet to be explored. This article proposes a machine learning-based approach to address this challenge, aiming to optimize both the supply and manufacturing planning phases as a practical solution for industry planning or optimization problems. The proposed approach examines supervised machine learning and deep learning techniques for manufacturing time and cost planning in various scenarios of a large-scale real-life pilot study in the bicycle manufacturing domain. This experimentation included K-Nearest Neighbors with regression and Random Forest from the machine learning family, as well as Neural Networks and Ensembles as deep learning approaches. Additionally, Reinforcement Learning was used in scenarios where real-world data or historical experiences were unavailable. The training performance of the pilot study was evaluated using cross-validation along with two statistical analysis methods: the t-test and the Wilcoxon test. These performance evaluation efforts revealed that machine learning techniques outperform deep learning methods and the reinforcement learning approach, with K-NN combined with regression yielding the best results. The proposed approach was validated by industry experts in bicycle manufacturing. It demonstrated up to a 37% reduction in both time and cost for orders compared to traditional expert estimates. Full article

This paper explores the application of machine learning techniques and statistical analysis to identify the patterns of victimization and the risk of gender-based violence in San Andrés de Tumaco, Nariño, Colombia. Models were developed to classify women according to their vulnerability and risk of suffering various forms of violence, which were integrated into a decision-making tool for local authorities. The algorithms employed include K-means for clustering, artificial neural networks, random forests, decision trees, and multiclass classification algorithms combined with fuzzy classification techniques to handle the incomplete data. Implemented in Python and R, the models were statistically validated to ensure their reliability. Analysis based on health data revealed the key victimization patterns and risks associated with gender-based violence in the region. This study presents a data science model that uses a social determinant approach to assess the characteristics and patterns of violence against women in the Pacific region of Nariño. This research was conducted within the framework of the Orquídeas Program of the Colombian Ministry of Science, Technology, and Innovation. Full article