• 대한원격탐사학회지
• /
• 제38권6_4호
• /
• pp.1911-1923
• /
• 2022

The availability of high-resolution satellite images provides precise information without direct observation of the research target. Korea Multi-Purpose Satellite (KOMPSAT), also known as the Arirang satellite, has been developed and utilized for earth observation. The machine learning model was continuously proven as a good classifier in classifying remotely sensed images. This study aimed to compare the performance of the support vector machine (SVM) model in classifying the land cover of the Delaware River port area on high and medium-resolution images. Three optical images, which are KOMPSAT-2, KOMPSAT-3A, and Sentinel-2B, were classified into six land cover classes, including water, road, vegetation, building, vacant, and shadow. The KOMPSAT images are provided by Korea Aerospace Research Institute (KARI), and the Sentinel-2B image was provided by the European Space Agency (ESA). The training samples were manually digitized for each land cover class and considered the reference image. The predicted images were compared to the actual data to obtain the accuracy assessment using a confusion matrix analysis. In addition, the time-consuming training and classifying were recorded to evaluate the model performance. The results showed that the KOMPSAT-3A image has the highest overall accuracy and followed by KOMPSAT-2 and Sentinel-2B results. On the contrary, the model took a long time to classify the higher-resolution image compared to the lower resolution. For that reason, we can conclude that the SVM model performed better in the higher resolution image with the consequence of the longer time-consuming training and classifying data. Thus, this finding might provide consideration for related researchers when selecting satellite imagery for effective and accurate image classification.

• 한국컴퓨터정보학회논문지
• /
• 제27권11호
• /
• pp.123-130
• /
• 2022

최근 API Call 정보를 기반으로 안드로이드 악성코드를 탐지 및 분류하는 연구가 활발하게 진행되고 있다. 그러나 API Call 기반의 악성코드 분류는 방대한 데이터 양과 높은 차원 특성으로 인해 악성코드 분석과 학습 모델 구축 과정에서 과도한 시간과 자원이 소모된다는 심각한 제한사항을 가진다. 이에 본 연구에서는 방대한 API Call 정보를 포함하고 있는 CICAndMal2020 데이터세트를 대상으로 PCA(Principal Component Analysis, 주성분분석)를 사용하여 차원을 대폭 축소시킨 후 LightGBM, Random Forest, k-Nearest Neighbors 등의 다양한 분류 기법 모델을 적용하여 결과를 분석하였다. 그 결과 PCA가 원본 데이터의 특성을 유지하면서 데이터 특성의 차원은 획기적으로 감소시키고 우수한 악성코드 분류 성능을 달성함을 확인하였다. 이진분류 및 다중분류 모두 데이터 특성을 전체 크기의 1% 수준 이하로 줄이더라도 이전 연구 결과보다 높은 수준의 정확도를 나타내었다.

• Wind and Structures
• /
• 제36권4호
• /
• pp.277-292
• /
• 2023

Low-rise structures are generally immersed within the roughness layer of the atmospheric boundary layer flows and represent the largest class of the structures for which wind loads for design are being obtained from the wind standards codes of distinct nations. For low-rise buildings, wind loads are one of the decisive loads when designing a roof. For the case of cylindrical roof structures, the information related to wind pressure coefficient is limited to a single span only. In contrast, for multi-span roofs, the information is not available. In this research, the numerical simulation has been done using ANSYS CFX to determine wind pressure distribution on the roof of low-rise cylindrical structures arranged in rectangular plan with variable spacing in accordance with building width (B=0.2 m) i.e., zero, 0.5B, B, 1.5B and 2B subjected to different wind incidence angles varying from 0° to 90° having the interval of 15°. The wind pressure (P) and pressure coefficients (C_pe) are varying with respect to wind incidence angle and variable spacing. The results of present numerical investigation or wind induced pressure are presented in the form of pressure contours generated by Ansys CFD Post for isolated as well as variable spacing model of cylindrical roofs. It was noted that the effect of wind shielding was reducing on the roofs by increasing spacing between the buildings. The variation pf Coefficient of wind pressure (C_pe) for all the roofs have been presented individually in the form of graphs with respect to angle of attacks of wind (AoA) and variable spacing. The critical outcomes of the present study will be so much beneficial to structural design engineers during the analysis and designing of low-rise buildings with cylindrical roofs in an isolated as well as group formation.

• Nuclear Engineering and Technology
• /
• 제55권9호
• /
• pp.3409-3416
• /
• 2023

Licensing the next-generation of nuclear reactor designs requires extensive use of Modeling and Simulation (M&S) to investigate system response to many operational conditions, identify possible accidental scenarios and predict their evolution to undesirable consequences that are to be prevented or mitigated via the deployment of adequate safety barriers. Deep Learning (DL) and Artificial Intelligence (AI) can support M&S computationally by providing surrogates of the complex multi-physics high-fidelity models used for design. However, DL and AI are, generally, low-fidelity 'black-box' models that do not assure any structure based on physical laws and constraints, and may, thus, lack interpretability and accuracy of the results. This poses limitations on their credibility and doubts about their adoption for the safety assessment and licensing of novel reactor designs. In this regard, Physics Informed Neural Networks (PINNs) are receiving growing attention for their ability to integrate fundamental physics laws and domain knowledge in the neural networks, thus assuring credible generalization capabilities and credible predictions. This paper presents the use of PINNs as surrogate models for accidental scenarios simulation in Nuclear Power Plants (NPPs). A case study of a Loss of Heat Sink (LOHS) accidental scenario in a Nuclear Battery (NB), a unique class of transportable, plug-and-play microreactors, is considered. A PINN is developed and compared with a Deep Neural Network (DNN). The results show the advantages of PINNs in providing accurate solutions, avoiding overfitting, underfitting and intrinsically ensuring physics-consistent results.

• 응용통계연구
• /
• 제35권6호
• /
• pp.739-754
• /
• 2022

텍스트 데이터는 일반적으로 많은 변수를 포함하고 있으며 변수들 사이의 연관성도 높아 통계 분석의 정확성, 효율성 등에서 문제가 생길 수 있다. 이러한 문제점에 대처하기 위해 목표 변수가 주어진 지도 학습에서는 목표 변수를 잘 설명할 수 있는 단어들을 선택하여 이 단어들만 통계 분석에 이용하기도 한다. 반면, 비지도 학습에서는 목표 변수가 주어지지 않으므로 지도 학습에서와 같은 단어 선택 절차를 활용하기 어렵다. 이 연구에서는 토픽 모형을 이용하여 지도 학습에서의 목표 변수를 대신할 수 있는 토픽을 생성하고 각 토픽별로 연관성이 높은 단어들을 선택하는 단어 선택 절차를 제안한다. 제안된 절차를 실제 텍스트 데이터에 적용한 결과, 단어 선택 절차를 이용하면 많은 토픽에서 공통적으로 자주 등장하는 단어들을 제거함으로써 토픽을 더 명확하게 식별할 수 있었다. 또한, 군집 분석에 적용한 결과, 군집과 범주 사이에 높은 연관성을 가지는 군집 분석 결과를 얻을 수 있는 것으로 나타났다. 목표 변수에 대한 정보없이 토픽 모형을 이용하여 선택한 단어들을 분류 분석에 적용하였을 때 목표 변수를 이용하여 단어들을 선택한 경우와 비슷한 분류 정확성을 얻을 수 있음도 확인하였다.

• 융합보안논문지
• /
• 제23권1호
• /
• pp.117-123
• /
• 2023

점점 더 고도화되고 있는 랜섬웨어 공격을 기계학습 기반 모델로 탐지하기 위해서는, 분류 모델이 고차원의 특성을 가지는 학습데이터를 훈련해야 한다. 그리고 이 경우 '차원의 저주' 현상이 발생하기 쉽다. 따라서 차원의 저주 현상을 회피하면서 학습모델의 정확성을 높이고 실행 속도를 향상하기 위해 특성의 차원 축소가 반드시 선행되어야 한다. 본 논문에서는 특성의 차원이 극단적으로 다른 2종의 데이터세트를 대상으로 3종의 기계학습 모델과 2종의 특성 추출기법을 적용하여 랜섬웨어 분류를 수행하였다. 실험 결과, 이진 분류에서는 특성 차원 축소기법이 성능 향상에 큰 영향을 미치지 않았으며, 다중 분류에서도 데이터세트의 특성 차원이 작을 경우에는 동일하였다. 그러나 학습데이터가 고차원의 특성을 가지는 상황에서 다중 분류를 시도했을 경우 LDA(Linear Discriminant Analysis)가 우수한 성능을 나타냈다.

• Fisheries and Aquatic Sciences
• /
• 제26권6호
• /
• pp.380-392
• /
• 2023

Groupers (Family Epinephelidae) are commonly caught in data-poor small-scale multi-species fisheries for sale on both export and domestic markets. This study presents data on the species composition and size/life-stage structure of Epinephelus spp. groupers caught by small-scale fishers and sold locally in the Indonesian province of South Sulawesi. Data were collected from fishing ports and local markets at 12 sites representing the three seaways around South Sulawesi (Makassar Strait, Flores Sea, Gulf of Bone). Each specimen (n = 3,398) was photographed alongside an object of known length, and total length (TL) was obtained using the Rapid Scaling on Object (RASIO). Of the 23 species identified, four (Epinephelus areolatus, Epinephelus ongus, Epinephelus quoyanus, and Epinephelus fasciatus) collectively comprised 69% of the catch, while the 13 least abundant species contributed less than 5%. The catch was dominated (67%) by the subadult life-stage, with just under 20% in the adult class. Juveniles dominated the catch of Epinephelus fuscoguttatus, a valuable export commodity. Observations of early maturity as well as the sizeable gap between length at first capture (Lc) and length at first maturity (Lm) indicate recruitment overfishing of most species, with the notable exception of Epinephelus rivulatus. The proportion of adult fish was low (≈5%-30%) for the twelve most abundant species (E. areolatus, E. ongus, Epinephelus quoyanus, E. fasciatus, Epinephelus coioides, Epinephelus faveatus, Epinephelus sexfasciatus, Epinephelus maculatus, Epinephelus bleekeri, Epinephelus corallicola, E. fuscoguttatus, Epinephelus polyphekadion). For two moderately abundant species (E. faveatus and E. malabaricus), TL < Lm for all specimens. The limited data available indicate spawning ratio is lower than reported from deep-water fisheries of E. areolatus and E. coioides. The results call for targeted research to fill knowledge gaps regarding the biology and ecology of groupers exploited mainly for domestic markets; highlight the need for species-level data to inform management policies such as minimum legal size regulations; and can contribute towards species-level status assessments.

• 영어영문학
• /
• 제56권5호
• /
• pp.973-998
• /
• 2010

This paper aims at assessing the significance of Lyrical Ballads of 1798, the agreed inaugurator of English Romanticism, in terms of such key concepts as poetic "experiments," "conversation," and the authorial function. The 1798 volume marks an interesting incidence in which an author with no tangible substantiality can wield his authorial function over his works. The volume is signed without the named proper noun-its author is neither William Wordsworth nor Samuel Taylor Coleridge. The figure of the author in this case is realized by the poems he writes; he produces, and is produced by, his works-a fact that constitutes part of the poetic experiments manifested in the Advertisement. Working under this reciprocal production, the Author of the 1798 volume and his poems are collectively aiming at establishing a new class of poetry and an interpretive community. The notion of "conversation" is a key element in the thematic, stylistic ties among individual poems. Poems of the 1798 volume effect multi-layered, "blended" voices. Readers are expected to draw out the topological interweaving among poems through the practices of dialogic reading. In this light, the sequential necessity of "The Rime" and "Tintern Abbey" should be emphasized. They are stitched together in a logic of textual placement and the transition from one to the other is never arbitrary. Most of all, they are working under the same authorial function, complementing each other, and addressing the same poetic project in different textual locations. As an inaugural work of English Romanticism, Lyrical Ballads of 1798 in fact makes so many things happen and yet again anticipates something yet to come with elusiveness. The value of this poetic experiments should be judged not only by what is claimed in it, but what it sets out to do and "how far" it will be performed, as implied in the Advertisement. The efficacy of the volume, more than anything else, is dependent upon the performative power of words.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2022년도 제66차 하계학술대회논문집 30권2호
• /
• pp.81-84
• /
• 2022

제조 산업에서의 이상치 검출은 생산품의 품질과 운영비용을 절감하기 위한 중요한 요소로 최근 딥러닝을 사용하여 자동화되고 있다. 이상치 검출을 위한 딥러닝 기법에는 CNN이 있으며, CNN을 계층적으로 구성할 경우 단일 CNN 모델에 비해 상대적으로 성능의 향상을 보일 수 있다는 것이 많은 선행 연구에서 나타났다. 이에 MVTec-AD 데이터셋을 이용하여 계층 CNN이 다중 클래스 이상치 판별 문제에 대해 효과적인지를 탐구하고자 하였다. 실험 결과 단일 CNN의 정확도는 0.7715, 계층 CNN의 정확도는 0.7838로 다중 클래스 이상치 판별 문제에 있어 계층 CNN 방식 접근이 다중 클래스 이상치 탐지 문제에서 알고리즘의 성능을 향상할 수 있음을 확인할 수 있었다. 계층 CNN은 모델과 파라미터의 개수와 리소스의 사용이 단일 CNN에 비하여 기하급수적으로 증가한다는 단점이 존재한다. 이에 계층 CNN의 장점을 유지하며 사용 리소스를 절약하고자 하였고 K-means, GMM, 계층적 클러스터링 알고리즘을 통해 제작한 새로운 클래스를 이용해 계층 CNN을 구성하여 각각 정확도 0.7930, 0.7891, 0.7936의 결과를 얻을 수 있었다. 이를 통해 Clustering 알고리즘을 사용하여 적절히 물체를 분류할 경우 물체에 따른 개별 상태 판단 모델을 제작하는 것과 비슷하거나 더 좋은 성능을 내며 리소스 사용을 줄일 수 있음을 확인할 수 있었다.

• 한국작물학회:학술대회논문집
• /
• 한국작물학회 2022년도 추계학술대회
• /
• pp.88-88
• /
• 2022

Detection of stress responses in crops is important to diagnose crop growth and evaluate yield. Also, the multi-spectral sensor is effectively known to evaluate stress caused by nutrient and moisture in crops or biological agents such as weeds or diseases. Therefore, in this experiment, multispectral images were taken by an unmanned aerial vehicle(UAV) under field condition. The experiment was conducted in the long-term fertilizer field in the National Institute of Crop Science, and experiment area was divided into different status of NPK(Control, N-deficiency, P-deficiency, K-deficiency, Non-fertilizer). Total 11 vegetation indices were created with RGB and NIR reflectance values using python. Variations in nutrient content in plants affect the amount of light reflected or absorbed for each wavelength band. Therefore, the objective of this experiment was to evaluate vegetation indices derived from multispectral reflectance data as input into machine learning algorithm for the classification of nutritional deficiency in rice. RandomForest model was used as a representative ensemble model, and parameters were adjusted through hyperparameter tuning such as RandomSearchCV. As a result, training accuracy was 0.95 and test accuracy was 0.80, and IPCA, NDRE, and EVI were included in the top three indices for feature importance. Also, precision, recall, and f1-score, which are indicators for evaluating the performance of the classification model, showed a distribution of 0.7-0.9 for each class.