Online-Journals.org (International Association of Online Engineering)
Not a member yet
9805 research outputs found
Sort by
Evaluating the Effectiveness of Remote Mobile Health Services for the Elderly: Standards and Best Practices
Get PDFThe innovative application of emerging mobile and wearable health information and sensing technologies (mHealth) holds great promise for lowering healthcare costs and enhancing well-being in various ways. Numerous fields are seeing the development of these applications. Still, to fully understand the opportunities and difficulties associated with using mobile technologies to enhance health outcomes, more thorough research is required. There is currently little data supporting the effectiveness of mHealth. Even though these technologies seem harmless and enticing, more research is required to determine the best times, locations, and users for mHealth methods, applications, and devices. As a result, a remote medical hub waitperson is essential to sustainably providing high-quality telemedicine services. This paper offers a thorough analysis of the delivery of medical services through telemedicine applications, with a focus on medical center servers. It also draws attention to the problems and obstacles that still need to be resolved to deliver health services via telehealth in the medical center. These ideas are used in this work to define a high standard for the developing field of mHealth research, explore potential future directions, and describe existing assessment standards
General Super-Resolution Techniques: A Literature Review
Get PDFSuper-resolution (SR) is a technique aimed at improving the resolution of images. In blood cell imaging, it aids in the accurate identification and classification of cells. Improving the analysis process of microscopic images is necessary to achieve better disease diagnoses, especially the image quality, so that health professionals can reach a diagnosis closer to the ideal. For those aiming to implement SR algorithms to analyze microscopic blood cell images, it is crucial to determine which algorithms are in use, their intended purposes, future trends, and current gaps. No review of SR techniques focusing on blood cells was found in the literature. Therefore, this paper presents various techniques to improve the resolution of blood images. Data screening and inclusion followed the PRISMA method. Articles were grouped into four subtopics: generic (25.0%), vascular imaging (28.1%), cell imaging (9.4%), and blood cell imaging techniques (37.5%). Results revealed that more research efforts on cell imaging techniques would be required to achieve a more balanced distribution. This study contributes to knowledge by reviewing the most used techniques, their purposes, and applications, helping researchers find the best technique for their studies, especially for pathological researchers involved in image enhancement
AI-Enhanced Biosignal Analysis for Obstructive Sleep Apnea Detection: A Comprehensive Review
Get PDFObstructive sleep apnea (OSA) detection using single-lead electrocardiograms (ECGs) has advanced significantly with the integration of artificial intelligence (AI). This review explores how AI enhances feature extraction and machine learning algorithms to improve OSA detection. The RR interval in electrocardiographic data is particularly valued for its ease of identification and low error rate. We review a range of machine learning and deep learning techniques employed in OSA detection. This review offers insights into developing single-lead ECG-based OSA detection systems by analyzing database availability, feature extraction methods, and machine learning approaches
ECG Biometric Authentication Using Deep CNN Feature Learning from Analytic Wavelet-Transformed Signals
Get PDFThis paper investigates the use of continuous morse wavelet transform (CWT) coefficients as inputs to convolutional neural networks (CNNs) for electrocardiogram (ECG) biometric authentication. We evaluate the performance and generalization of pre-trained SqueezeNet architecture using the ECG-ID Database. Our approach involves extracting 10 scalograms from each subject’s ECG signals and employing gradient descent optimization during training. The models demonstrate high accuracy, achieving over 90% on both training and validation datasets, indicating robust performance and minimal overfitting. Further analysis using the F1 confidence curve and ROC curve reveals a balanced trade-off between precision and recall, with an optimal F1 score of 0.84 and an AUC of 0.84, respectively. Additionally, we explore the impact of different CWT parameter settings, including Voice per Octave (VPO), symmetry parameter (gamma), and time-bandwidth product (P2). The optimal VPO of 41 yields an AUC of 0.87 and an F1 score of 0.84. The best performance is achieved with gamma values greater than 2 and time-bandwidth products between 45 and 80, enhancing time localization and frequency resolution. In this study, the significance of fine-tuning wavelet parameters to improve the effectiveness of ECG biometric systems is demonstrated, demonstrating the potential of combining CWT and CNNs for reliable biometric authentication
An Intrusion Detection Model for Internet of Medical Things Using BDA-DAN2 Model
Get PDFThe Internet of Medical Things (IoMT) is a subset of the Internet of Things (IoT) where medical devices communicate with one another to share sensitive data. The integration of medical devices into the IoT has greatly assisted the development of the IoMT. These advancements facilitate effective communication and providing care for patients in the healthcare sector. However, they also face specific security and privacy concerns, such as malware attacks and denial of service (DoS) attacks. To overcome this problem, intrusion detection systems (IDS) are introduced, specifically employing deep learning (DL) methodologies. This study proposes a deep learning-based binary dragonfly algorithm (BDA) with a dynamic architecture for artificial neural networks2 (DAN2) model for implementing a robust and accurate IDS in IoMT. The IDS has the following stages: collection of data, preprocessing, selection of features, and classification. The IoMT dataset is employed to train the model to get improved outcomes. The standard scalar technique is used for the data preprocessing process. The BDA algorithm is used for feature selection (FS) of the preprocessed data. The DAN2 model is implemented to classify the selected data and to improve the classification accuracy. The dataset was further divided for training and testing of the model. The performance of the BDA-DAN2 model is assessed utilizing the evaluation parameters of accuracy, recall, precision, and F1-score. The BDA-DAN2 model demonstrates superior performance with 99.12% accuracy, 99.28% precision, 99.40% recall, and 98.56% F1-score during training, and 98.92% accuracy, 98.50% precision, 98.68% recall, and 97.90% F1-score during testing. Experiments confirmed that the binary dragonfly algorithm with the DAN2 (BDA-DAN2) model has the highest accuracy compared to the existing models
Hybrid Approach Using Multi-Relational Weighted Matrix Factorization (WMRMF) and Cohen’s Kappa (Sk) to Refine Educational Items Clustering
Get PDFIn the context of adopting the competency-based approach (CBA) as a new teaching methodology in sub-Saharan countries, and particularly in Côte d’Ivoire, the development of learning content that is aligned with the economic, socio-cultural, and scientific needs of society is of paramount importance. Educational experts have therefore proposed competencies and associated tasks for educational programs. However, these learning contents often face issues of task redundancy. The present paper aims to address this problem by proposing a hybrid approach to educational item clustering, combining weighted multi-relational matrix factorization (WMRMF) and Cohen’s Kappa (Sk) techniques (Sk-WMRMF). This approach takes into account not only student performance and achievements but also a novel reflexive relationship, “tasks-require–tasks”. To evaluate the Sk-WMRMF approach, we conducted a survey among students in general secondary schools in Côte d’Ivoire. With an accepted task redundancy threshold of 0.185, an RMSE score of 0.198, and an improvement rate of 90.47% in the “tasks–skills” mapping, the results demonstrate that Sk-WMRMF enhances the elaboration of “task-skill” mappings. This not only improves learning content but also facilitates the updating of curricula in accordance with the CBA approach
The Impact of ChatGPT on English Language Learners' Writing Skills: An Assessment of AI Feedback on Mobile
Get PDFArtificial intelligence (AI) has shown promise in enhancing English as a second language (ESL) writing skills by providing personalized feedback and targeted corrections, thereby facilitating improved grammar and composition proficiency. Despite the potential of AI tools like ChatGPT, their impact on common writing errors in ESL contexts has yet to be explored. This study employed a quasi-experimental design to compare the efficacy of ChatGPT’s mobile application feedback against traditional teacher feedback in a senior secondary public school in India. Over eight weeks, the experimental group received feedback on their writing error corrections through the ChatGPT application, while the control group received feedback from teachers. Additionally, participants’ attitudes towards using ChatGPT for language learning were assessed through a questionnaire administered post-intervention to 132 students. Data was collected using pre- and post-tests that involved writing stories based on pictures. The study results demonstrated that the experimental group significantly improved writing proficiency, showing a reduction in common errors (third-person singular present, past tense, progressive, past participle, plural, possessive, comparative, and superlative) compared to the control group. Furthermore, most students preferred AI feedback, associating it with noticeable improvements in their writing skills and grammatical accuracy. These findings support the integration of AI tools like ChatGPT into language learning curricula as effective supplements to traditional teaching methods, offering personalized and immediate corrections that enhance learning outcomes
Application of Smart Mobile Devices in Electronic Design Education: Multidimensional Interaction Model and Learning Outcomes Assessment
Get PDFWith the rapid advancement of technology, smart mobile devices are increasingly being integrated into the educational domain, showing significant potential, particularly in electronic design education. Traditional classroom teaching methods face limitations in information delivery and student interaction, while the introduction of smart mobile devices brings new opportunities to classroom instruction. Through smart mobile devices, educators can organize teaching activities more flexibly, and students can engage in classroom interactions in various forms, greatly enhancing teaching effectiveness and learning experiences. Although numerous studies have explored the application of smart mobile devices in education, most focus on single-dimensional interaction models, overlooking the potential of multidimensional interactions. Additionally, traditional methods for assessing learning outcomes often rely on qualitative analysis and post-class tests, which fail to comprehensively and in realtime reflect students’ learning states and emotional changes. This paper aims to construct a multidimensional interaction model for electronic design education classrooms based on smart mobile devices and to assess learning outcomes through real-time analysis of students’ emotions and feedback data, thereby optimizing teaching strategies. This study not only provides new perspectives and methods for the application of smart mobile devices in education but also offers practical guidance for teaching reforms and innovations in electronic design education, holding significant theoretical and practical value
Analyzing the Correlation Between Student Learning Behaviors and Psychological Atmosphere Using Deep Learning
Get PDFWith the rapid development of educational technology, the application of deep learning in analyzing student behavior and psychological states has become a hot topic in the field of education. This study, grounded in deep learning technology, aims to explore the correlation between student learning behaviors and psychological atmosphere, as well as the positive impact of a constructive psychological atmosphere on student learning. The background section discusses the limitations of traditional educational assessment methods and the necessity and urgency of applying deep learning in education. The current state of study section presents the progress in analyzing the correlation between student behavior and psychological states, highlighting the shortcomings in data processing and model construction in existing studies. Addressing these shortcomings, this study proposes a student learning behavior detection scheme based on the lightweight neural network shufflenetV2 and validates through empirical study the positive influence of a constructive psychological atmosphere on student learning behavior. The results show that utilizing a lightweight network model can effectively identify patterns in student learning behavior and, to some extent, predict the students’ psychological states. Furthermore, a positive psychological atmosphere indeed enhances students’ learning motivation and behavior. The methods and findings of this study hold significant theoretical and practical implications for advancing personalized and intelligent education
Using a Random Controlled Trial to Explore the Impact of AI-Enabled Learning
Get PDFMuch is being made of AI’s role in learning. However, there have been few studies that evaluate AI’s ability to make learning more efficient or more effective and specifically the learner’s willingness to embrace and use AI-enabled learning tools over conventional learning methodologies.
Working with a manufacturing company, we used a randomized control trial approach (matching pairs) to present two groups of employees with a course on design thinking, one presented via a traditional learning management system (LMS Group) and other through an AI enabled tool (AI Group), with the content being identical in order to compare completion rates, time to completion, learning outcomes (subject matter knowledge retained) and the learner’s perceptions of value and experience. We also were able to compare the AI enabled version of the course to the same version with the addition of a face-to-face instructional component.
While we found that those participants in the AI Group completed the course in a shorter period of time, had better learning outcomes and expressed higher perceptions of course value, the AI Group significantly underperformed the LMS Group in terms of course commencement and completion rates. Our finding suggest that the adoption of AI enabled learning tools may follow similar patterns of adoption rates of new technologies -- and require educating learners regarding the use, benefits, value and limitations of new AI enabled tools before widespread acceptance and usage