Earticle

Many ESL and EFL learners find wh-question formation difficult to master despite the fact that wh-interrogatives are a commonly used structure. Through using a Learning Study approach, this study aims to identify Hong Kong ESL learners’ difficulties with wh-question formation, to explore effective strategies to enhance the instruction, and to investigate the learning outcomes as a result. Quantitative data were collected from three Hong Kong Secondary 3 classes through a written test and an oral test, and qualitative data from discussion at teachers’ meetings, research lesson observation, student interviews, and post-lesson conferences. It was found that Hong Kong learners’ L1 transfer problems are an important inhibiting factor in their learning of English wh-interrogatives. The most serious learning problems found include Chinese interference with word order and failure to use correct verb phrase structures. The research team, through the Learning Study research cycle, succeeded in designing appropriate contexts and using variation patterns for presenting and practicing interrogative structures. Successful outcomes were evident in the gradual improvement of the design of the three research lessons and significant progress by students in their learning of wh-interrogatives.

Background Abdominal obesity measures such as waist circumference and waist-to-hip ratio have been shown to relate more closely to diabetes and musculoskeletal pain than body mass index, but conventional tape-based measurements are not easily scalable. A smartphone inclinometer application provides a practical alternative by capturing abdominal tilt angles that are automatically computed and easily reusable. Purpose This study assessed whether abdominal tilt angles measured using a smartphone inclinometer could serve as digital features for classifying diabetes and musculoskeletal pain in older adults using machine learning. Methods In 105 older adults, 12 abdominal inclination features were extracted from upper and lower tilt angles and refined by minimum redundancy maximum relevance selection. Three models (Light Gradient Boosting Machine (LightGBM), Balanced Random Forest (BalancedRF), and Linear Support Vector Classifier with Probability Calibration) were trained and evaluated with five-fold cross-validation. Results BalancedRF achieved the best performance for diabetes classification (accuracy = 0.83, ROC-AUC = 0.93, PR-AUC = 0.84), with sensitivity 87% and specificity 82%. For musculoskeletal pain, LightGBM achieved moderate performance (accuracy = 0.78, ROC-AUC = 0.83, PR-AUC = 0.56) but sensitivity was limited (53%) despite high specificity (87%). SHAP analysis highlighted the squared terms of the lower and total abdominal angles as key features for diabetes, while associations were weaker for pain. Conclusions Abdominal tilt angles measured by a smartphone inclinometer represent feasible, noninvasive digital features for diabetes risk stratification, although utility for musculoskeletal pain classification is limited. Future work should validate these findings in larger and longitudinal cohorts and explore real-world integration of smartphone posture monitoring for digital health applications.

Deep learning addresses a wide range of complex challenges, spanning from computer vision to data analytics. It is also employed to develop softwares that pose threats to privacy and security. To develop a DeepFake video, an individual in the original video is replaced with someone else using deep learning. Various deep learning-based techniques have been proposed to detect DeepFakes. In this work, we extensively analyse DeepFake video detection techniques considering their strengths and limitations. We provide a comparative analysis along with discussing their architectures and performances. Finally, we propose hyperparameter settings that improve deep learning model’s overall accuracy and efficiency.

Regression analysis is one of the most widely utilized methods because of its adaptability and simplicity. Recently, the machine learning (ML) approach, which is one aspect of regression methods, has been gaining attention from researchers, including social science, but there are only a few studies that compared the traditional approaches with the ML approach. This study was conducted to explore the usefulness of the ML approach by comparing the ordinary least square estimate (OLS), the stochastic gradient descent algorithm (SGD), and the support vector regression (SVR) with a model predicting and explaining the tuberculosis screening intention. The optimized models were evaluated by four aspects: computational speed, effect and importance of individual predictor, and model performance. The result demonstrated that each model yielded a similar direction of effect and importance in each predictor, and the SVR with the radial kernel had the finest model performance compared to its computational speed. Finally, this study discussed the usefulness and attentive points of the ML approach when a researcher utilizes it in the field of communication.

Cardio-Vascular Diseases (CVD) is found to be rampant in the populace leading to fatal death. The statistics of a recent survey reports that the mortality rate is expanding due to obesity, cholesterol, high blood pressure and usage of tobacco among the people. The severity of the disease is piling up due to the above factors. Studying about the variations of these factors and their impact on CVD is the demand of the hour. This necessitates the usage of modern techniques to identify the disease at its outset and to aid a markdown in the mortality rate. Artificial Intelligence and Data Mining domains have a research scope with their enormous techniques that would aassist in the prediction of the CVD priory and identify their behavioural patterns in the large volume of data. The results of these predictions will help the clinicians in decision making and early diagnosis, which would reduce the risk of patients becoming fatal. This paper compares and reports the various Classification, Data Mining, Machine Learning, Deep Learning models that are used for prediction of the Cardio-Vascular diseases. The survey is organized as threefold: Classification and Data Mining Techniques for CVD, Machine Learning Models for CVD and Deep Learning Models for CVD prediction. The performance metrics used for reporting the accuracy, the dataset used for prediction and classification, and the tools used for each category of these techniques are also compiled and reported in this survey.

On-device machine learning is fundamentally constrained by limited storage, especially in continuous data collection scenarios where sensor or vision streams accumulate rapidly. This paper empirically investigates storage-aware learning, characterizing the trade-off between data quantity and data quality under lossy compression. Using the CIFAR-10 dataset as a controlled benchmark, we systematically vary both the amount and the fidelity of training data to understand their joint impact on model performance. Our results reveal that (1) neither maximizing quantity nor quality alone yields optimal accuracy, emphasizing that the optimal trade-off between them depends nonlinearly on the available storage budget, and (2) data samples exhibit differential sensitivity to compression, motivating a sample-wise adaptive compression policy. These findings challenge uniform data-retention strategies such as naive data dropping or fixed-rate compression, and establish a foundation for adaptive, storage-efficient learning systems on resource-limited devices. This work opens new directions toward generalizable, storage-aware on-device intelligence.

Purpose: Nursing students' competence in virtual simulation-based learning is a key factor in its success. This study explored the validity and reliability of a virtual-simulation-based learning competency self-evaluation tool for nursing students. Methods: Data were collected from a web-based survey. First, 11 nursing professors participated in a focus group interview, and 7 simulation education experts participated in the preliminary item content validity. The participants in these two aspects were not the same. Then, a preliminary survey was conducted with 15 fourth-year nursing students in I City. Next, based on these three efforts, a final survey comprising 20 evaluation items was developed. This survey was administered to third- and fourth-year nursing students at four nursing colleges in Korean provinces (Seoul, Gyeonggi, Gangwon, and Gyeongsan-do); 222 complete questionnaires were used for the final analysis. Further, Kirkpatrick’s evaluation model was used for four steps each of tool development and verification processes of the associated psychometric aspects, for a total of eight steps. An exploratory factor analysis was performed on the collected survey data, and verify the tool's validity and reliability. Results: Four factors comprising 15 items explained 66.59% of the variance: learning preparation and start-up (4 items), nursing assessment (3 items), data interpretation (3 items), and problem solving (5 items). The Cronbach's α of the tool was 0.74, and that of the factors ranged from 0.72 to 0.80. Conclusions: The tool's validity and reliability were demonstrated using established methodologies. This tool can be useful for evaluating Korean nursing students' virtual simulation learning competence.

Purpose: This exploratory, qualitative study was conducted to explore Chinese couples’ perceived educational needs for parenting a child with specific learning difficulties (SLD). Methods: We recruited a purposive sample of 25 couples who were caring for a child with SLD at home from one regional child mental health center in Hong Kong. Data were collected from individual couples via audio-taped, semi-structured interviews between April and June 2011. Each interview lasted for about an hour. We analyzed the interview data using qualitative content analysis, as suggested by Graneheim and Lundman (2004). Results: Four identified categories of parents’ perceived educational needs were information needs for caregiving, a variety of health concerns with themselves and their child, inadequate psychosocial support, and perceived stigma attached to help-seeking. These results reveal challenges and information for mental health professionals in providing effective educational and psychosocial support and culturespecific health care for these families and their children with learning disabilities. Conclusion: Our findings indicate a few important educational needs of parents in caring for a child with SLD that might be underestimated by mental health professionals and teachers, such as psychological support and information needs. To facilitate effective parenting, holistic and individualized needs assessment and education should be provided to address each parent’s biopsychosocial and cultural needs in relation to caregiving.

Purpose: This study aimed to develop a predictive model for the early identification of patients at risk of sepsis, using routinely available clinical information and laboratory test results collected during the initial phase of patient care. Methods: This retrospective analysis included electronic medical records of 22,400 adult patients who presented with suspected infection to a tertiary care university hospital in Korea between January 2013 and May 2024. Patients were classified according to Systemic Inflammatory Response Syndrome (score ≥ 2) or Quick Sequential Organ Failure Assessment (score ≥ 2), in combination with sepsis-related International Classification of Diseases, 10th revision codes. Four different machine learning models were trained and validated using five-fold cross-validation. In addition, Shapley additive explanations analysis was performed to interpret the contribution and clinical relevance of key predictive variables. Results: Among the evaluated models, CatBoost demonstrated the strongest predictive performance. Notably, platelet distribution width, alveolar?arterial oxygen difference, procalcitonin, and the arterial/alveolar oxygen ratio consistently emerged as major predictors. Importantly, several variables that did not reach statistical significance in univariate analysis nevertheless contributed substantially to overall model performance, highlighting the importance of complex, multidimensional interactions among clinical factors. Conclusion: These findings indicate that a model based on simple, routinely collected clinical data can achieve high predictive accuracy and strong generalizability. Such a tool may support early clinical decision-making by multidisciplinary teams, including nurses, across diverse real-world care settings. Further prospective studies are warranted to validate its clinical utility and to assess its potential effects on patient outcomes.

Purpose: The purposes of this study were to develop a prediction model for pressure injury using a machine learning algorithm and to integrate it into clinical practice. Methods: This was a retrospective study of tertiary hospitals in Seoul, Korea. It analyzed patients in 12 departments where many pressure injuries occurred, including 8 general wards and 4 intensive care units from January 2018 to May 2022. In total, 182 variables were included in the model development. A pressure injury prediction model was developed using the gradient boosting algorithm, logistic regression, and decision tree methods, and it was compared to the Braden scale. Results: Among the 1,389,660 general ward cases, there were 451 cases of pressure injuries, and among 139,897 intensive care unit cases, there were 297 cases of pressure injuries. Among the tested prediction models, the gradient boosting algorithm showed the highest predictive performance. The area under the receiver operating characteristic curve of the gradient boosting algorithm's pressure injury prediction model in the general ward and intensive care unit was 0.86 (95% confidence interval, 0.83~0.89) and 0.83 (95% confidence interval, 0.79~0.87), respectively. This model was integrated into the electronic health record system to show each patient's probability for pressure injury occurrence, and the risk factors calculated every hour. Conclusion: The prediction model developed using the gradient boosting algorithm exhibited higher performance than the Braden scale. A clinical decision support system that automatically assesses pressure injury risk allows nurses to focus on patients at high risk for pressure injuries without increasing their workload.

Purpose: This study aimed to develop and compare machine learning models for predicting prediabetes in young adults in Korea using dietary intake data and to identify the most effective model. Methods: Data from the ninth Korea National Health and Nutrition Examination Survey were used, with 823 participants aged 19-35 years selected after excluding those with missing data. Logistic regression, k-nearest neighbors, and random forest models were applied to predict prediabetes, and the analysis was conducted using the Orange 3.5 program. Five-fold cross-validation was performed to reduce performance variability, and test data were used for final model validation. Results: In the dataset, 14%-15% of participants were classified as having prediabetes. The random forest model showed the highest performance in terms of classification accuracy, harmonic mean of precision and recall, and precision. Logistic regression had the highest performance regarding the model’s ability to distinguish between individuals with and without prediabetes. Age, thiamine intake, and water intake emerged as the most important predictors. Conclusion: This study demonstrated the utility of using dietary intake data to predict prediabetes in young adults. The random forest model provided the highest prediction accuracy, supporting early detection and intervention, which could help to reduce unnecessary treatment. This highlights nurses’ important role in educating patients about lifestyle changes and implementing preventive care. Future studies should incorporate additional factors, such as psychological and lifestyle variables, to improve the model's performance.

Purpose: This study aimed to seek agreement on the core components of an e-learning cultural competence program for Korean public health workers (PHWs) while prioritizing educational content areas. Methods: A two-round Delphi study was performed with 16 Korean experts from five disciplines. Data were collected between August 30 and November 24, 2017. A questionnaire was developed from literature reviews and previous focus group interviews concerning PHWs. The panel members were asked to rate the importance and urgency of educational content areas and the effectiveness of teachingelearning methods and gave opinions on their appropriate frequency, duration, and target audience. Responses were analyzed using descriptive statistics. A median of 4.0 or greater or a rate of agreement of 75% or greater was considered a “consensus” for the purposes of this study. Results: All consenting participants responded to two-round surveys. Participants reached consensus on thirty-one educational content areas. Of these, the highest priorities were “necessity of cultural competence of PHWs,” “health characteristics according to race and ethnicity,” and “establishing trusting relationships with migrants.” The most effective teachingelearning method was case-based learning, with seven to eight sessions of training and duration of fewer than 30 minutes per session. Conclusion: Multidisciplinary experts proposed 12 prioritized educational content areas and effective teachingelearning methods as well as their frequency, duration, and target audiences, reflecting Koreaspecific multicultural phenomena and the nature of the work of PHWs. These findings can contribute to preparing PHWs to provide culturally competent services to migrants in their communities.

Purpose: Clinical reasoning, which is based on an understanding of the pathophysiological mechanismsof diseases, is a core nursing competency that involves analyzing patient-related data and providingappropriate nursing practices. Simulation-based education is effective in improving clinical reasoningcompetencies and communication skills. This study evaluated the effectiveness of virtual simulation-basedlearning in improving the communication skills and clinical reasoning competencies ofundergraduate nursing students. Methods: This study used a single-group pretest and posttest quasi-experimentaldesign to evaluate the effectiveness of virtual simulation-based learning. Data werecollected from June to September 2020. Thirty-seven nursing students in their third and fourth yearsof study who understood the purpose of this study were selected as participants. The collected datawere analyzed using SPSS Statistics 25.0 and Winsteps 3.68.2. Results: The communication skills (t =?12.80, p < .001) and clinical reasoning competency (t = ?4.67, p < .001) of the undergraduate nursingstudents who participated in the virtual simulation-based learning program improved significantlyafter participation. Additionally, a Rasch model analysis revealed that the overall clinical reasoningcompetency of undergraduate nursing students improved. Conclusion: Virtual simulation-based learningprograms for nursing students should be developed and implemented.

본 연구는 ‘좌우뇌 활용 능력 향상을 위한 한자 교육 콘텐츠 개발’의 후속 연구로, 개발한 콘텐츠에 대한 전 문가 평가로 마무리된 연구에 이어서 사용자 관점에서 문제점을 발견하고 이를 보완함으로써 교육 콘텐츠로 서의 학습 가이드라인을 제시하는 것을 목적으로 한다. 1차 사용자 평가는 사용자 관찰법을 통해 진행하여 발견된 문제에 대하여 보완하고, 2차 사용자 평가에서는 참여자에 대한 뇌성향 검사를 진행하고, 인지면접법을 통해 사용자 평가를 진행하였다.

This study is a follow-up study on the development of East Asian Character education content to improve the ability to utilize the left and right brain. Following the study, which was completed with an expert evaluation of the developed content, problems were discovered and supplemented from the user's perspective. Through this, the purpose is to present learning guidelines as educational content. The first user evaluation was conducted through user observation to correct for discovered problems, and in the second user evaluation, brain propensity tests were conducted on participants, and user evaluation was conducted through cognitive interviewing. In this study, the following conclusions can be presented. First, it should be recognized that some educational approaches may need to be approached differently for left-brained and right-brained users. Second, questions that take a visual approach should be designed so that only one answer is correct. Third, for questions that take a linguistic approach, the answers are presented so that the differences due to contrast are clearly visible, and the answer is approached by looking for one point that is different from the standard East Asian Character. Fourth, it is appropriate to select East Asian Characters that will be used as learning objects in content from among those who are skeptical. Fifth, right-brained users most preferred the type that understood meaning from visually forced images, while left-brained users most preferred the type that understood meaning by combining elements.

In this paper, “the valuable diagnostic tool” Strategy Inventory for Language Learning (SILL), proposed by Oxford (1990), has been employed to conduct an empirical study on learning strategies of tertiary-level EFL learners in China on the basis of a brief survey of the research into L2 learning strategies. The paper examines and analyzes the overall frequency with which they employ learning strategies. The relationship between learning strategies and three other variables (gender, time spent in English learning, and English language proficiency) is also examined through the well-known SPSS 10.0. The findings of the research suggest that the learning strategies employed most frequently by the tertiary-level EFL learners are memory, cognitive and metacognitive strategies, followed by compensation strategies, with social and affective strategies the least frequently used. The study also shows that significant difference exists between the employment of learning strategies and gender in English learning, and that strategy use demonstrates a significant relationship, to varying degrees, with the time spent on English learning and English proficiency as well. Finally, the paper discusses the causes of such a relationship between the above-mentioned variables, and points out the major factors that affect the strategy use of the tertiary-level EFL learners in China.

본 연구는 대학 신입생을 대상으로 한 파이썬 입문 강의에서 블록 기반, 텍스트 기반, 혼합형 교수법의 효과성을 조사하였다. 60명의 학생이 세 그룹으로 무작위 배정되어 각기 다른 교수법을 13주간 적용받았다. 학습 성과, 문제 해결 능력, 학생 태도는 시험, 코딩 연습, 설문 조사 및 인터뷰를 통해 평가되었다. 결과적으로 모든 그룹에서 유의미한 향상이 나타났으며, 혼합형 교수법을 적용한 그룹이 가장 높은 사후 평가 점수(81.2점)와 최종 프로젝트에서 최고 성과(평균 89.3점)를 기록하였다. 혼합형 교수법은 블록 기반 프로그래밍의 접근성과 텍스트 기반 코딩의 심층성을 효과적으로 결합하여, 시각적 표현에서 텍스트 표현으로의 점진적인 전환을 통해 학습을 단계적으로 지원하였다. 이 접근법은 다양한 학습자의 요구를 충족시키면서도 고급 프로그래밍 기술 습득을 준비하도록 설계되었다. 혼합형 그룹의 학생들은 자신감, 즐거움, 몰입감에서 가장 큰 향상을 보고하였다. 연구 결과는 혼합형 교수법이 개념적 이해, 코딩 숙련도 및 학습 만족도를 증진하는 데 가장 효과적임을 강조하고 있다.

This study investigates the effectiveness of block-based, text-based, and hybrid instructional approaches in introductory Python courses for community college freshmen. Sixty students were randomly assigned to three groups, each following a distinct instructional method over a 13-week semester. Learning outcomes, problem-solving skills, and student attitudes were evaluated using exams, coding exercises, surveys, and interviews. The results revealed significant improvements across all groups, with the hybrid approach achieving the highest post-test scores (81.2 points) and the best project performance(average of 89.3 points in the final project). The hybrid method effectively combined the accessibility of block-based programming with the depth of text-based coding, employing a gradual transition from visual to textual representations. This approach scaffolded learning, addressing the needs of diverse learners while equipping them with advanced programming skills. Students in the hybrid group reported the greatest improvements in confidence, enjoyment, and engagement. The findings underscore the hybrid approach as the most effective for promoting conceptual understanding, coding proficiency, and overall satisfaction among diverse learners.

The aim of this study is (1) to examine the underlying constructs of grammar learning strategies identified by Korean EFL learners and (2) to investigate the relationship between grammar strategy and grammar/vocabulary attainment of Korean English learners divided by English proficiency levels. The quantitative method was employed, using factor analysis and a series of ANOVA analysis based on the data collected through a strategy use survey, a grammaticality judgement test and a vocabulary levels test. It involved 65 participants that took a TOEIC class at a university as part of regular course requirements for a semester. The research finding showed that the underlying three factors of the instrument were not identical with the factors produced by Korean EFL learners. Instead, the five factors were emerged from the current study. The results of a series of ANOVA exhibited that there partially existed statistical differences among the proficiency levels in terms of Factor 3 (Noticing & awareness raising) for the grammaticality judgement test with Factor 4 (self-directed grammar learning) showing significance for the vocabulary levels test. The high proficiency level learners more often utilized various types of GLS while low level learners were shown to use less GLS. Some considerations and directions for further research are suggested.

기존의 얕은 인공신경망은 과적합과 Gradient Vanishing 현상 등 내재적 문제점으로 인하여 산업시설 의 화재 위험성 평가에 한계가 있었다. 그러나 최근 들어 은닉층을 다층으로 구성하는 깊은 신경망 의 구축이 가능해지고 학습알고리즘이 고도화되면서 화재보험에서 화재위험성 평가도구로서 활용 성이 높아졌다. 본 논문에서는 구글사의 텐서플로우를 이용하여 다양한 학습조건에서 깊은 신경망 을 학습시켜 얕은 신경망의 예측성능과 비교하였다. 그 결과 깊은 신경망에서는 Drop-out 및 ReLU함 수의 활용을 통해 기존의 SNN의 문제점을 해소할 수 있었으며 TS값이 최대 0.76으로 얕은 신경망보 다 58% 높은 학습성능을 확인하였다. 그러나 화재보험에서 위험관리도구로서 활용성을 높이기 위해 서는 체계적이고 많은 데이터가 확보되어야 한다.

The shallow learning neural network (SNN) has some limitations in the assessment of fire risk of industrial facilities due to its inherent problems such as over-fitting and gradient vanishing. However, in recent years, it has become possible to build a deep learning neural network (DNN) consisting of multiple hidden layers and to make learning algorithms more sophisticated, which allows for the use of a fire risk assessment tool in the fire insurance. In this paper, prediction performances between SNN and DNN are compared under various conditions using Google's Tensorflow. As a result, most SNN problems are solved through the drop-out method and ReLU activation function in DNN, and the learning performance of DNN with a maximum TS value of 0.76 is confirmed to be 58% higher than that of SNN. Nevertheless, in order to improve the utilization of fire insurance as a risk management tool, a systematic and large amount of learning data should be secured.