• Nuclear Engineering and Technology
• /
• 제52권10호
• /
• pp.2262-2273
• /
• 2020

Digital twin technology can provide significant value for the prognostics and health management (PHM) of critical plant components by improving insight into system design and operating conditions. Digital twinning of systems can be utilized for anomaly detection, diagnosis and the estimation of the system's remaining useful life in order to optimize operations and maintenance processes in a nuclear plant. In this regard, a conceptual framework for the application of digital twin technology for the prognosis of Control Element Drive Mechanism (CEDM), and a data-driven approach to anomaly detection using coil current profile are presented in this study. Health management of plant components can capitalize on the data and signals that are already recorded as part of the monitored parameters of the plant's instrumentation and control systems. This work is focused on the development of machine learning algorithm and workflow for the analysis of the CEDM using the recorded coil current data. The workflow involves features extraction from the coil-current profile and consequently performing both clustering and classification algorithms. This approach provides an opportunity for health monitoring in support of condition-based predictive maintenance optimization and in the development of the CEDM digital twin model for improved plant safety and availability.

• Nuclear Engineering and Technology
• /
• 제52권5호
• /
• pp.1008-1012
• /
• 2020

In order to make reasonable design for the improvement of comprehensive mechanical properties of RAFM steels, the design system with both machine learning and high-throughput optimization algorithm was established. As the basis of the design system, a dataset of RAFM steels was compiled from previous literatures. Then, feature engineering guided random forests regressors were trained by the dataset and NSGA II algorithm were used for the selection of the optimal solutions from the large-scale solution set with nine composition features and two treatment processing features. The selected optimal solutions by this design system showed prospective mechanical properties, which was also consistent with the physical metallurgy theory. This efficiency design mode could give the enlightenment for the design of other metal structural materials with the requirement of multi-properties.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제31권8호
• /
• pp.1083-1091
• /
• 2004

본 논문은 기계번역 시스템에서 파싱의 전처리 단계로 도입되는 구 묶음 시스템을 제안한다. 구 묶음 모듈은 구 묶음의 결과로 얻어지는 의존관계 제약을 통하여 분석 시스템의 성능향상에 기여할 수 있다. 중국어를 위한 구 묶음 시스템을 변형 기반 학습 기법에 근거하여 구현하며, 의존관계를 효과적으로 파서에 넘겨줄 수 있는 인터페이스를 고안한다. 구현된 모듈을 중한 기계번역 시스템에 통합하고, 중국 관련 웹사이트로부터 수집한 말뭉치를 이용한 실험을 통해 구 묶음의 도입이 기계번역에서 분석시스템의 성능 향상에 기여할 수 있음을 보인다.

• Journal of the Korean Data and Information Science Society
• /
• 제27권5호
• /
• pp.1273-1284
• /
• 2016

본 연구에서는 기계 학습 분야의 특허를 수집하여 키워드 네트워크를 구축하고 클릭 분석을 실시하였다. 먼저 텍스트 마이닝 기법을 적용하여 핵심 키워드들을 선정한 다음, 이 키워드를 기반으로 키워드 네트워크를 구축하였다. 다음으로 네트워크 구조 분석, 중요 키워드 분석 및 클릭 분석을 시행하여 2005년도와 2015년도에 출원된 기계 학습 특허의 동향을 파악하였을 뿐만 아니라 양해년도의 분석 결과를 통해 특허 경향을 파악하였다. 분석 결과 기계 학습 특허의 키워드 네트워크는 밀도와 군집 계수가 낮은 것으로 드러났으며 기계 학습 기법 자체에 대한 특허보다는 다양한 응용 영역에서 기계학습을 적용한 특허들이 다수이기 때문으로 판단된다. 클릭 분석 결과 2005년도 클릭 분석에 의해 발견된 주제는 뉴스메이커 검증, 상품 소비 예측, 바이러스 공격 예방, 바이오마커, 그리고 워크플로우 관리였으며, 2015년도 기계 학습 특허 주제는 디지털 이미지 편집, 직불카드, 수신자 인라이닝 시스템, 유방 촬영 시스템, 재고 관리 시스템, 이미지 편집 시스템, 비행기 티켓 가격 예측, 그리고 문제 예측 시스템으로 나타났다. 2005년도에 비하여 2015년도의 근접 중앙성은 낮아지고 매개 중심성은 높아진 것으로 보아 최근의 특허 경향은 보다 다양한 분야에서 출원되고 있으며 이들 간의 연결이 활발해지고 있음을 알 수 있다. 클릭 분석은 클릭을 형성하는 키워드 집합을 해석하여 주제를 파악하는데 활용될 수 있을 뿐만 아니라 추출된 공유 멤버쉽 키워드 집합은 특허 검색 시스템과 같이 키워드 검색 기반의 시스템에서 검색 키워드로 활용될 수 있을 것으로 기대된다.

• 정보보호학회논문지
• /
• 제28권3호
• /
• pp.635-642
• /
• 2018

4차 산업혁명 시대에 우리는 소프트웨어 홍수 속에 살고 있다. 그러나, 소프트웨어의 증가는 필연적으로 소프트웨어 취약점 증가로 이어지고 있어 소프트웨어 취약점을 탐지 및 제거하는 작업이 중요하게 되었다. 현재까지 소프트웨어 취약 여부를 예측하는 연구가 진행되었지만, 탐지 시간이 오래 걸리거나, 예측 정확도가 높지 않았다. 따라서 본 논문에서는 기계학습 알고리즘을 이용하여 소프트웨어의 취약 여부를 효율적으로 예측하는 방법을 설명하며, 다양한 기계학습 알고리즘을 이용한 실험 결과를 비교한다. 실험 결과 k-Nearest Neighbors 예측 모델이 가장 높은 예측률을 보였다.

• Journal of Multimedia Information System
• /
• 제5권3호
• /
• pp.195-200
• /
• 2018

Recently, as the sensor and big data analysis technology have been developed, there have been a lot of researches that analyze the purchase-related data such as the trajectory information and the stay time. Such purchase-related data is usefully used for the purchase pattern prediction and the purchase time prediction. Because it is difficult to find periodic patterns in large-scale human data, it is necessary to look at actual data sets, find various feature patterns, and then apply a machine learning algorithm appropriate to the pattern and purpose. Although existing papers have been used to analyze data using various machine learning methods, there is a lack of statistical analysis such as finding feature patterns before applying the machine learning algorithm. Therefore, we analyze the purchasing data of Songjeong Maeil Market, which is a data gathering place, and finds some characteristic patterns through statistical data analysis. Based on the results of 1, we derive meaningful conclusions by applying the machine learning algorithm and present future research directions. Through the data analysis, it was confirmed that the number of visits was different according to the regional characteristics around Songjeong Maeil Market, and the distribution of time spent by consumers could be grasped.

• 전기전자학회논문지
• /
• 제23권4호
• /
• pp.1400-1407
• /
• 2019

본 논문은 실시간 자가학습과 분류 기능을 모두 지원하는 support vector machine (SVM) 기반 가스 분류기의 하드웨어 구조 설계 및 구현 결과를 제시한다. 제안된 가스 분류기는 학습 알고리즘으로 modified sequential minimal optimization(MSMO)을 사용하였고, 학습과 분류 기능을 공유구조를 사용하여 설계함으로써 기존 논문 대비 하드웨어 면적을 35% 감소시켰다. 설계된 가스 분류기는 Xilinx Zynq UltraScale+ FPGA를 사용하여 구현 및 검증되었고, 108MHz의 동작 주파수에서 3,337개의 CLB LUTs로 구현 가능함을 확인하였다.

• 한국물환경학회지
• /
• 제37권5호
• /
• pp.335-343
• /
• 2021

Water quality prediction is essential for the proper management of water supply systems. Increased suspended sediment concentration (SSC) has various effects on water supply systems such as increased treatment cost and consequently, there have been various efforts to develop a model for predicting SSC. However, SSC is affected by both the natural and anthropogenic environment, making it challenging to predict SSC. Recently, advanced machine learning models have increasingly been used for water quality prediction. This study developed an ensemble machine learning model to predict SSC using the XGBoost (XGB) algorithm. The observed discharge (Q) and SSC in two fields monitoring stations were used to develop the model. The input variables were clustered in two groups with low and high ranges of Q using the k-means clustering algorithm. Then each group of data was separately used to optimize XGB (Model 1). The model performance was compared with that of the XGB model using the entire data (Model 2). The models were evaluated by mean squared error-ob servation standard deviation ratio (RSR) and root mean squared error. The RSR were 0.51 and 0.57 in the two monitoring stations for Model 2, respectively, while the model performance improved to RSR 0.46 and 0.55, respectively, for Model 1.

• 한국BIM학회 논문집
• /
• 제12권2호
• /
• pp.1-11
• /
• 2022

Road traffic accidents (RTAs) are the leading cause of accidental death among children. RTA reduction is becoming an increasingly important social issue among children. Municipalities aim to resolve this issue by introducing "Smart Pedestrian Crosswalks" that help prevent traffic accidents near children's facilities. Nonetheless such facilities tend to be installed in relatively limited number of areas, such as the school zone. In order for budget allocation to be efficient and policy effects maximized, optimal location selection based on machine learning is needed. In this paper, we employ machine learning models to select the optimal locations for smart pedestrian crosswalks to reduce the RTAs of children. This study develops an optimal location index using variable importance measures. By using k-means clustering method, the authors classified the crosswalks into three types after the optimal location selection. This study has broadened the scope of research in relation to smart crosswalks and traffic safety. Also, the study serves as a unique contribution by integrating policy design decisions based on public and open data.

• ETRI Journal
• /
• 제45권1호
• /
• pp.105-118
• /
• 2023

We tested the feasibility of automated discrimination of patients with panic disorder (PD) from healthy controls (HCs) based on multimodal physiological responses using machine learning. Electrocardiogram (ECG), electrodermal activity (EDA), respiration (RESP), and peripheral temperature (PT) of the participants were measured during three experimental phases: rest, stress, and recovery. Eleven physiological features were extracted from each phase and used as input data. Logistic regression (LoR), k-nearest neighbor (KNN), support vector machine (SVM), random forest (RF), and multilayer perceptron (MLP) algorithms were implemented with nested cross-validation. Linear regression analysis showed that ECG and PT features obtained in the stress and recovery phases were significant predictors of PD. We achieved the highest accuracy (75.61%) with MLP using all 33 features. With the exception of MLP, applying the significant predictors led to a higher accuracy than using 24 ECG features. These results suggest that combining multimodal physiological signals measured during various states of autonomic arousal has the potential to differentiate patients with PD from HCs.