한국산 미기록종인 붉은광택긴가슴잎벌레(신칭)를 처음으로 보고한다. 유충에 대한 자세한 형태학적 설명과 그림, COI 유전자 서열, 성충 사진, 그리고 생식기 그림을 제공한다. 또한 긴가슴잎벌레속 성충에 대한 분류학적 키를 제공한다.

Catch per unit effort (CPUE) is widely utilized as an index of stock abundance in fisheries assessments, but its interpretation is often complicated by temporal and spatial variation in fishing activity. For this reason, CPUE standardization is essentially needed to produce indices that better reflect the actual stock status. In Korea, however, the limited availability of detailed operational data has posed challenges for effective CPUE standardization. In this study, CPUE data for sailfin sandfish (Arctoscopus japonicus) caught by the East Sea mid-sized Danish seine fishery were standardized using records from the Korean Fishery Radio Station (FRS) from 2004 to 2024. The dataset mainly consists of fishing dates, locations (30’ × 30’ latitude-longitude grids), and catch weights. A generalized linear model (GLM) was applied, incorporating year, quarter, area, and their interactions as explanatory variables. Among these, the year was identified as the most influential factor, followed by quarter and area. The standardized CPUE showed a more stable trend than the nominal CPUE, which showed an increase from 2004 to 2017 and a sharp decline thereafter. This suggests that the standardized CPUE more accurately reflects the underlying dynamics of sailfin sandfish stock abundance. Despite the absence of detailed logbook records, the FRS records effectively captured the spatial distribution of fishing activity and operational patterns, demonstrating their potential as alternative data sources for CPUE standardization. Nevertheless, limitations remain due to the self-reported nature of the data, which may lead to low coverage and potential reliability issues. Improvements in both the quantity and quality of data collection and reporting are necessary to enhance the utility of such data in stock assessments. This study highlights the potential and challenges of using FRS data for CPUE standardization and provides practical guidance for its application in fisheries management.

This study aims to enhance accessibility in transportation-disadvantaged areas by utilizing Large Language Model(LLM) to analyze public transportation and advanced mobility status data (e.g., platform taxis and Demand Responsive Transport(DRT)), and proposes a methodology to support region-specific mobility activation strategies. The study was divided into three stages: first, the collection of mobility data; second, the implementation of geographic information system (GIS)-based visualization and preprocessing; and third, the application of LLM-based image interpretation and classification. A variety of mobility data were consolidated into a unified spatial entity, converted into visualization information for LLM processing, and examined using a rule-based classification system to ascertain the mobility environment types. This approach addresses the limitations of single-data analysis and enables a multi-layered interpretation of regional transportation gaps. Through the LLM interpretation of visual elements, including grid colors, patterns, bus routes, and designated DRT operation areas, transportation characteristics such as mobility supply levels, DRT operation status, and taxi dependency were identified. The LLM model demonstrated a high level of performance with a precision rate of 78.2 %, accuracy rate of 73.1 %, recall rate of 91.8 %, and F1-score of 84.5 %. Notably, the recall rate exceeded 90 %, signifying comprehensive recognition of various transportation environment types. This study proposes an LLM-based spatial data interpretation framework for analyzing regional mobility conditions in Paju City. The integration of complex spatial information into QGIS enables the LLM to automatically analyze data, thereby unveiling micro-level mobility characteristics and identifying four types of regional mobility improvements.

최근 BIM은 단순 3차원 모델링을 넘어 표준화된 속성 데이터의 품질 확보와 체계적 관리가 핵심 요구로 부각되고 있으며, 온톨로 지 및 지식그래프 기반의 데이터 관리･추론 방식이 주목받고 있다. 그러나 지식그래프 기반 BIM 데이터는 부재 간 구조 관계 분석, 설 계 검토, 물량･속성 정보 조회 등 실무 의사결정을 지원할 수 있으나, SPARQL･Cypher와 같은 그래프 질의어를 직접 작성해야 한다 는 점에서 실무 적용에 제약으로 작용한다. 이를 위해 본 연구에서는 사용자가 자연어 질문을 기반으로 그래프 질의를 자동 생성할 수 있는 GraphRAG 기반 질의 자동 생성 프레임워크를 제안하였다. 먼저 CSV 기반 속성/관계 테이블에 규칙을 적용해 노드･관계를 생 성하고 그래프 데이터베이스에 적재하는 CSV-to-LPG 파이프라인을 구현하여, LPG 지식그래프 구축 절차를 자동화하였다. 이후 Few-shot Learning 기반 프롬프트 설계를 통해 사용자의 자연어 질문을 Cypher 쿼리로 자동 변환하는 자동 질의 생성 모듈을 구현하 였다. 전체 프레임워크는 Graph-ACQ 시스템으로 개발하여 라멘교 BIM 데이터를 기반으로 적용하였다. 검증 결과 LPG 스키마 유효 성과 Cypher 자동 생성, Cypher 질의 수작업 과정에서 정확도 모두 100%를 달성하였고, 질의 생성 시간은 평균 7.1초에 처리되었다. GraphRAG 기반 질의 생성 방식은 부재 간 공간･구조 관계를 명시적으로 활용하므로, 설계･검토 과정에서 요구되는 연결 관계 분석, 구조 구성 파악, 물량･속성 정보 조회 등 관계 기반 질의를 자연어로 수행할 수 있다. 또한 Few-shot Learning 기반 접근을 적용하여 교 량 뿐만 아니라 다양한 공종 내에서도 질의 생성을 가능하게 함으로써, 프로젝트의 확장성을 확보 가능하다.

This study develops a scientific fishing-ground exploration framework for the Korean large purse-seine fishery, where traditional experience-based searching has become increasingly unreliable under rapid climate variability. AIS-derived fishing locations from 2021 to 2023 were integrated with HYCOM-based temperature and salinity fields and MODIS-Aqua chlorophyll-a data to construct a unified environmental – fishing dataset. After multicollinearity screening and principal component analysis, temperature and salinity at 30 m depth and chlorophyll-a were selected as representative predictors. Using these variables, a generalized additive model (GAM) with background-sampled pseudo-absence data and monthly maximum entropy (MaxEnt) models were developed to quantify nonlinear habitat – environment relationships and predict monthly and seasonal mackerel fishing occurrences. Model performance was evaluated using independent data from 2024. GAM exhibited relatively stable predictive performance across months with generally high AUC and TSS values whereas MaxEnt showed pronounced seasonal variability and was effective in identifying potential habitat structures based on presence-only environmental conditions. Spatial predictions from both models showed good agreement with observed fishing-ground distributions during specific seasons, reproducing high-suitability zones associated with seasonal thermal – salinity fronts and productivity gradients. These results provide insights into the environmental mechanisms governing purse-seine fishing grounds and demonstrate the complementary roles of GAM for operational prediction and MaxEnt for potential habitat exploration.

This study investigated the role of snack consumption in daily nutrient intake among Korean adults. Data from The Korea National Health and Nutrition Examination Survey (KNHANES) 2023 were analyzed, involving 4,918 adults, consuming 2,118 men (43.1%) and 2,800 women (56.9%). The analysis assessed daily energy intake from snacks and its percentage of total energy, categorized by sex, age, income, and BMI, using multiple regression to explore associations. On average, 18.8% of daily energy intake, with men at 17.3% and women at 20.7%. Women had higher energy contributions for major nutrients. Snack energy contributions were significantly influenced by sex and age, but not by income or BMI. Young adults (ages 19—29) had the highest contribution at 21.1%, while adults aged 50—64 had the lowest at 17.5%. Time-of-day analysis revealed that 22.7% of snack energy was consumed in the evening (18:00—20:59) and 22.6% at night (21:00 onward). The most commonly consumed snacks were coffee (23.8%) and fruit (20.3%). These findings emphasize that snacks are a significant source of energy and nutrients, highlighting the need for sexand age-specific nutritional interventions that consider snack type, timing, and quality.

The purpose of this study is to analyze the recent incidence rates of Behçet's disease using big data. Big data for this study was used for academic purposes in the 2019 to 2022 rare disease patient statistics from the Korea Disease Control and Prevention Agency (public domain source). A total of 3,886 newly diagnosed patients with Behçet's disease in Koreans from 2019 to 2022 were analyzed. In this study, the incidence of Behçet's disease was analyzed according to gender and age, and the incidence in the total population was also analyzed. The incidence rates of Behçet's disease from 2019 to 2022 were compared and analyzed at 1-year intervals. The incidence rate in men was in the 30% range, and in women, it was in the 60% range. The incidence of Behçet's disease in individuals over their 50s, approaching the elderly age group, was 42%. In the younger age group under 30, it was 16%. In this study, when analyzing the 4-year average incidence of Behçet's disease by age group from 2019 to 2022, it was found to be most common in people in their 40s at 24% over the past four years, followed by 50s at 22%, 30s at 17%, 60s at 14%, 20s at 12%, and 10s and 70s at 3%, showing a wide age distribution. This suggests that Behçet's disease can occur at any age.

Largehead hairtail (Trichiurus lepturus) is a commercially important warm-water species widely distributed in Korean coastal and offshore waters as well as adjacent waters. Recently, recreational fishing catches of this species have increased substantially, raising concerns about their contribution to total fishing pressure and the potential bias in stock assessment based solely on commercial catches. In this study, recreational fishing catches of largehead hairtail from 1970 to 2023 were reconstructed and combined with commercial catches to estimate total removals. Stock assessments were conducted using the CMSY (Catch-MSY) method under two scenarios: one using only commercial catches and the other incorporating recreational catches into the total catches. The results indicate that the current stock status of largehead hairtail is not overfished (B/BMSY > 1) and not subject to overfishing (F/FMSY < 1). However, the probability of overfishing increased compared to the scenario using only commercial catches when including recreational catches. This suggests that stock status may be overestimated if recreational removals are ignored. These findings highlight the importance of incorporating recreational fishing into stock assessments and indicate that systematic management of recreational fisheries should be implemented alongside commercial fisheries to ensure the sustainable use of the stock.

농업 분야 컴퓨터 비전(Computer Vision) 기술 확산으로 고품질 학습 데이터 확보가 필수적이나, 기존의 수동 데이터 구축 방식은 많은 시간과 비용이 소요되는 한계가 있다. 이에 본 연구는 최신 멀티모달 파운데이션 모델인 SAM3(Segment Anything Model 3)를 기반으로 반자동 어노테이션 시스템을 개발하였다. 제안 시스템은 (1) 텍스트 프롬프트 기반 객체 인 식, (2) SAM3 기반 정밀 마스크 생성 및 학습 가능한 폴리곤 좌표 변환, (3) 사용자 검증의 3단계로 구성되며 GUI로 구현 되었다. 600장 이미지 평가 결과, SAM3는 92.9%의 매칭률 과 0.790의 평균 정밀도(mAP)를 달성하였으며, 데이터셋 구 축 시간을 수동 작업 대비 96~98% 단축시켰다. 이는 SAM+ CLIP, Grounding DINO+SAM 등 기존 파운데이션 모델 대 비 정확도와 효율성 모든 면에서 월등한 성능이다. 본 연구는 파운데이션 모델의 제로샷 성능을 활용해 농업 데이터 레이블 링 효율을 개선하고 관련 AI 연구 가속화에 기여할 것으로 기 대된다.

Background: Real-time ergonomic risk assessment in manufacturing environments is challenged by severe class imbalance in high-risk postures and the need for deployment-efficient models. Conventional oversampling techniques may violate biomechanical constraints, limiting their suitability for human motion data. Objectives: This study aimed to compare multiple machine learning models for real-time ergonomic risk assessment while addressing data imbalance using biomechanically appropriate learning strategies and evaluating both predictive performance and deployment efficiency. Design: Comparative study. Methods: A large-scale workplace safety dataset comprising image-based skeletal keypoints was analyzed. To mitigate class imbalance without generating biomechanically implausible samples, cost-sensitive learning and focal loss were employed instead of synthetic oversampling. Subject-wise data splitting was applied to prevent data leakage. Five model families, including Random Forest, convolutional neural networks, and a lightweight graphbased network, were evaluated using accuracy, F1-score, area under the receiver operating characteristic curve (AUC), and high-risk recall. Statistical significance was assessed using bootstrap confidence intervals and McNemar and DeLong tests. Results: The lightweight graph-based model demonstrated competitive classification performance while maintaining reduced computational complexity. Although none of the models achieved the predefined high-risk recall threshold, statistically significant performance differences were observed across model families. Conclusion: The findings suggest that biomechanically informed imbalance handling improves methodological validity in ergonomic risk assessment. While deployment feasibility appears promising, further empirical validation on edge hardware is required.

This study addresses the challenge of imputing missing values in incomplete process data collected from high-cost data acquisition environments. Such missingness arises due to insufficient completeness, accuracy, and consistency, which significantly affect the quality of critical-to-quality (CTQ) attributes in manufacturing processes. We systematically evaluate three state-of-the-art imputation methods—Multiple Imputation by Chained Equations (MICE), the machine learning-based missForest algorithm, and a deep learning- based one-dimensional convolutional neural network (1D-CNN)—using real-world industrial data. Our analysis aims to identify the most effective imputation technique for handling complex and noisy process datasets typical in manufacturing settings. The results highlight the strengths and limitations of each method, providing practical guidance for selecting appropriate imputation approaches to improve the reliability of quality prediction and decision-making in industrial applications.

The Republic of Korea Armed Forces are implementing various scientific training systems to prepare for future warfare and are seeking advanced integration with AI - driven scientific military innovation. This study proposes the Synthetic Training Environment (STE) as a future-oriented model designed to overcome current limitations in military training. STE provides the foundation for evolving into an integrated, combined arms training system. In addition, the study introduces a data integration framework based on the Defense Training Management System (DTMS). This framework aims to standardize and unify training data across service branches, thereby enabling effective AI interoperability. Future tasks include real-time integration of synthetic and live training, 24/7 data access via the Metaverse, and the establishment of a cyclic system of learning, operation, and evolution. To that end, this research ultimately proposes the CJDSW-MST model - a comprehensive framework linking STE, DTMS, and unmanned combat systems for future-ready, intelligent military training.

The TDL (Tactical Data Link) network employs a TDMA (Time Division Multiple Access) scheme, in which the network resources of each wireless node are allocated to time slots. The method of time slot allocation for each node is based on expert experience and operational requirements. However, this method has limitations because it is difficult to verify real-world operational environments due to high costs and time requirements. To address these limitations, this study developed a TDMA simulator using SimPy, a Python-based discrete-event simulation framework. The proposed simulator enables analysis of time slot allocation methods under varying operational environment conditions. Simulation experiments were conducted to evaluate times slot requirements under different maximum message transmission delay time thresholds (6s and 12s). The results showed that stricter delay time thresholds and higher number of tracks increased the required number of time slots. In addition, the required number of time slots increased differently depending on the complex interaction of factors such as the number of tracks, delay time thresholds, operational scenarios. The proposed simulator provides more precise insights and supports more reliable TDL network design than conventional methods.

Anomaly detection is crucial for ensuring the reliability and safety of mechanical systems across industries such as power generation, manufacturing, and transportation. In these mechanical systems, data is usually collected in time-series form using sensors such like vibration, current or sound for anomaly detection. Time-series anomaly detection methods often face limitations due to insufficient training data and poor generalization across complex operating conditions and varying loads. To address these challenges, this study proposes a transfer learning-based anomaly detection model, leveraging pre-trained knowledge to deliver robust performance and adaptability in data-scarce scenarios and diverse industrial environments. To this end, time-series signals are transformed into spectrograms through Short-Time Fourier Transform(STFT), followed by feature extraction through a Convolutional Autoencoder to obtain low-dimensional latent features. These features are used to detect anomalies using classification such as Random Forest and eXtreme Gradient Boosting. Building on this approach, this research validates the model's performance through migration tasks using the Case Western Reserve University(CWRU) Bearings dataset. Furthermore, to show cross-condition generalization, the proposed model was validated on the Hanoi University of Science and Technology(HUST), Sumair–Umar Bearing Fault(SUBF) dataset v2.0, and a dataset collected using microphone sensor in motor dynamo tests. Consequently, unlike other studies limited by specific operating conditions, the proposed model exhibits strong generalization performance across benchmark datasets. Experimental results highlight the effectiveness of combining STFT, CAE, and tree-based classifiers in addressing data scarcity and enhancing generalization, making it highly suitable for real-world industrial applications. Future work will focus on noise-robust techniques and broader fault types to further improve performance.

With the rapid expansion of renewable energy deployment, power systems are increasingly exposed to issues such as higher output variability. Photovoltaic generation, as the most widely installed variable renewable energy source both domestically and internationally, exhibits significant fluctuations due to weather conditions. These characteristics lead to operational challenges including increased curtailment, higher reserve requirements, and even risks of large-scale outages. This study aimed to improve the accuracy of photovoltaic power generation forecasting by developing a data quality control procedure for meteorological data collected at a PV plant. The quality-controlled data were used as inputs to SVM and XGBoost, resulting in improved forecasting accuracy, with MAPE decreasing from 7–10% to 6.32% and 6.08%, respectively. The results demonstrate that meteorological data quality control significantly enhances PV forecasting performance and can contribute to distributed energy resource operation and curtailment mitigation strategies.

한류는 이제 단순한 문화 현상을 넘어, 세계적인 영향력을 가진 대한 민국의 주요한 관광자원으로 자리 잡았다. 본 연구는 소셜미디어를 통 해 수집한 한류관광 관련 키워드를 탐색 및 분석하고, 한류관광에 대한 활성화 방안을 제언하기 위해 진행되었다. 키워드 수집 기간은 2022년 9월 1일부터 2025년 8월 31일까지 약 3년간의 기간을 대상으로 수집 하였다. 분석 방법론은 빈도분석, 중심성 분석, CONCOR 분석을 적용 하여 단계별로 심층적인 분석을 진행하였다. 주요 연구 결과는 다음과 같다. 첫째, 한류관광과 연관성이 높은 키워드는 관광, 한류, 한국, 글 로벌, 팬덤, 패스, 문화, 콘텐츠 등이 확인되었다. 둘째, 한류관광 관련 연결중심성 분석을 시행한 결과, 관광, 한류, 문화, 콘텐츠, 팬덤, 산업 등의 키워드가 높은 수치를 보였으며, 근접중심성 분석에서는 문화, 외 국인, 여행, 세계 등의 키워드가 높은 수치를 나타냈다. 셋째, CONCOR 분석 결과, 총 3개의 군집이 도출되었다. 중심군집은 ‘K-컬 쳐 체험 및 지역관광’, ‘글로벌 K-팝 및 팬덤경제’이며, 주변군집은 ‘케 데헌 및 K-교통’으로 구성되었다. 도출된 3가지 군집을 대상으로 ‘촬영 지 및 공연지 기반의 지역관광 루트 개발’ 등의 정책적 한류관광 활성 화 방안을 구체적으로 제언하였다.

본 연구는 준거집단(취직자)들의 활동 데이터 뱅크를 생성하여 예비 취업자(고등교육기관의 체 육계열 전공자)들이 현재까지 활동했던 데이터를 데이터마이닝 기반 추천 알고리즘을 적용해 예비 취업자 들에게 가장 적합한 직업군을 추천해주는 스포츠 일자리 추천모형을 개발하고 검증하는 것이다. 따라서 평 가지표를 구성하고, 준거집단을 대상으로 인터뷰 및 조사를 통해 데이터 뱅크를 생성했다. 또한 비확률 표 본추출법 중 할당표본 추출법과 눈덩이표본 추출법을 적용해 예비 취업자 조사를 실시했으며, 총 921명의 자료를 통해 스포츠 일자리 추천모형 개발과 유사도를 통해 모형을 검증했다. 즉, 본 결과는 다음과 같다. 첫째, 준거집단과 예비 취업자의 평가지표를 구성했다. 둘째, 준거집단의 데이터 뱅크를 생성했다. 셋째, 스 포츠 전공 청년들을 위한 일자리 추천모형을 개발하고, 유사도를 통해 모형을 검증했다.

This study develops a post-value for money (post-VfM) evaluation model for public–private partnership (PPP) road projects in Korea. Following the abolition of the minimum revenue guarantee system, the demand risk was transferred to the private sector, thus necessitating an unbiased and data-driven assessment under the new adjusted build-transfer-operate (BTO-a) framework. The proposed model extends the existing ex-ante VfM analysis by incorporating actual operational data and estimating government payments for both public-sector comparator and private finance initiative alternatives on a lifecycle cost basis. Using an actual BTO project restructured as BTO-a, the simulation shows that the post-VfM ratio increases from 23.5% to 37.9%, thus confirming fiscal efficiency and balanced risk sharing. This model enables feedback between planning and operation, supports transparent policy evaluation, and provides a foundation for sustainable PPP governance in future infrastructure projects.

As renewable energy penetration continues to increase, the output variability and forecasting uncertainty of photovoltaic generation have emerged as major operational risks in power systems. This study establishes a sensor-based data quality control procedure to ensure the reliability of meteorological data collected at a PV plant. For temperature, humidity, and wind speed, a four stage QC process physical range check, persistence check, step change check, and median filtering was applied. Solar radiation, which exhibits strong temporal and distributional characteristics, was processed using a three-stage QC procedure consisting of physical range, step change, and frequency distribution checks. Using the quality-controlled meteorological data, PV generation forecasting was performed with SVM and XGBoost models. As a result, the MAPE values improved to 6.32% for SVM and 6.08% for XGBoost after QC application. The findings confirm that meteorological data quality control significantly enhances PV forecasting accuracy and can support future strategies for distributed energy resource management, curtailment mitigation, and power system risk reduction.

Truck platooning technology, which utilizes vehicle-to-vehicle communication to enable two or more autonomous trucks to travel in a platoon, is garnering attention. However, before platooning is implemented, an environment that can stably maintain a constant speed must be established. Therefore, maintaining a constant speed is a key prerequisite for truck platooning. To overcome the limitations of previous studies, which relied on traffic simulations or limited experiments, this study analyzes second-by-second truck DTG driving records obtained from highways near major domestic ports. Based on these data, a sliding-window technique was employed to detect constant-speed driving patterns and estimate the rate of constant-speed driving by section. The analysis revealed a high rate of constant-speed driving at the Noeun JCT–Dongcheongju IC, where the traffic volume was low and the road alignment was gentle. However, a low rate was observed at the Gunpo IC–Donggunpo IC, where ramp entries and exits were frequent. Subsequently, a multivariate fractional polynomial model was employed to analyze factors influencing constant-speed driving. The main factors identified were speed dispersion, average duration of constantspeed driving, and volume of large trucks per lane. This shows that speed stability, continuity of driving patterns, and vehicle composition within a section are more important factors in determining constant-speed driving than the average driving speed or traffic volume. This study is significant because it empirically elucidates the characteristics and factors influencing constant-speed driving using large-scale field data. Furthermore, it is expected to provide fundamental data for selecting suitable sections for truck platooning and establishing logistics efficiency policies.