• 한국데이타베이스학회:학술대회논문집
• /
• 한국데이타베이스학회 1999년도 춘계공동학술대회: 지식경영과 지식공학
• /
• pp.175-186
• /
• 1999

Detecting the features of significant patterns from their own historical data is so much crucial to good performance specially in time-series forecasting. Recently, a new data filtering method (or multi-scale decomposition) such as wavelet analysis is considered more useful for handling the time-series that contain strong quasi-cyclical components than other methods. The reason is that wavelet analysis theoretically makes much better local information according to different time intervals from the filtered data. Wavelets can process information effectively at different scales. This implies inherent support fer multiresolution analysis, which correlates with time series that exhibit self-similar behavior across different time scales. The specific local properties of wavelets can for example be particularly useful to describe signals with sharp spiky, discontinuous or fractal structure in financial markets based on chaos theory and also allows the removal of noise-dependent high frequencies, while conserving the signal bearing high frequency terms of the signal. To date, the existing studies related to wavelet analysis are increasingly being applied to many different fields. In this study, we focus on several wavelet thresholding criteria or techniques to support multi-signal decomposition methods for financial time series forecasting and apply to forecast Korean Won / U.S. Dollar currency market as a case study. One of the most important problems that has to be solved with the application of the filtering is the correct choice of the filter types and the filter parameters. If the threshold is too small or too large then the wavelet shrinkage estimator will tend to overfit or underfit the data. It is often selected arbitrarily or by adopting a certain theoretical or statistical criteria. Recently, new and versatile techniques have been introduced related to that problem. Our study is to analyze thresholding or filtering methods based on wavelet analysis that use multi-signal decomposition algorithms within the neural network architectures specially in complex financial markets. Secondly, through the comparison with different filtering techniques' results we introduce the present different filtering criteria of wavelet analysis to support the neural network learning optimization and analyze the critical issues related to the optimal filter design problems in wavelet analysis. That is, those issues include finding the optimal filter parameter to extract significant input features for the forecasting model. Finally, from existing theory or experimental viewpoint concerning the criteria of wavelets thresholding parameters we propose the design of the optimal wavelet for representing a given signal useful in forecasting models, specially a well known neural network models.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 1999년도 춘계공동학술대회-지식경영과 지식공학
• /
• pp.175-186
• /
• 1999

Detecting the features of significant patterns from their own historical data is so much crucial to good performance specially in time-series forecasting. Recently, a new data filtering method (or multi-scale decomposition) such as wavelet analysis is considered more useful for handling the time-series that contain strong quasi-cyclical components than other methods. The reason is that wavelet analysis theoretically makes much better local information according to different time intervals from the filtered data. Wavelets can process information effectively at different scales. This implies inherent support for multiresolution analysis, which correlates with time series that exhibit self-similar behavior across different time scales. The specific local properties of wavelets can for example be particularly useful to describe signals with sharp spiky, discontinuous or fractal structure in financial markets based on chaos theory and also allows the removal of noise-dependent high frequencies, while conserving the signal bearing high frequency terms of the signal. To data, the existing studies related to wavelet analysis are increasingly being applied to many different fields. In this study, we focus on several wavelet thresholding criteria or techniques to support multi-signal decomposition methods for financial time series forecasting and apply to forecast Korean Won / U.S. Dollar currency market as a case study. One of the most important problems that has to be solved with the application of the filtering is the correct choice of the filter types and the filter parameters. If the threshold is too small or too large then the wavelet shrinkage estimator will tend to overfit or underfit the data. It is often selected arbitrarily or by adopting a certain theoretical or statistical criteria. Recently, new and versatile techniques have been introduced related to that problem. Our study is to analyze thresholding or filtering methods based on wavelet analysis that use multi-signal decomposition algorithms within the neural network architectures specially in complex financial markets. Secondly, through the comparison with different filtering techniques results we introduce the present different filtering criteria of wavelet analysis to support the neural network learning optimization and analyze the critical issues related to the optimal filter design problems in wavelet analysis. That is, those issues include finding the optimal filter parameter to extract significant input features for the forecasting model. Finally, from existing theory or experimental viewpoint concerning the criteria of wavelets thresholding parameters we propose the design of the optimal wavelet for representing a given signal useful in forecasting models, specially a well known neural network models.

• 한국전산구조공학회논문집
• /
• 제34권1호
• /
• pp.51-58
• /
• 2021

본 논문에서는 Needle-punched C/SiC 복합재료 해석을 위한 효율적인 멀티스케일 해석기법을 소개한다. 기존 Needle-punching으로 인해 복잡한 미소구조를 갖는 NP 복합재료는 기존의 제안된 복합재료 멀티스케일 기법으로 물성을 계산하는 것은 한계가 있어 왔다. 이를 극복하기 위해 micro-CT 이미지 촬영을 통해 NP 복합재료의 미소구조를 면밀히 파악할 수 있었고, 이미지 프로세싱을 바탕으로 실제구조와 직접적으로 대응할 수 있는 3D high fidelity 모델을 구축하였다. 또한 유한요소해석에 맞춰 요소크기를 조절할 수 있는 sub-region processing 소개를 바탕으로 효율적인 유한요소해석을 수행하였다. NP 복합재료의 미소구조 거동뿐만 아니라, macro-scale 구조해석의 적용을 위해 subcell 모델링을 제안하였다. Needle-punching에 의한 Z축 NP 섬유의 규칙적인 간격을 이용하여 모델링을 수행할 수 있었다. 제안한 두 종류의 모델은 균질화 기법을 이용하여 등가거동 및 등가물성을 파악하였으며, 추가적인 실험 결과와의 비교를 통해 검증을 수행하였다.

• 한국제4기학회지
• /
• 제18권2호통권23호
• /
• pp.105-106
• /
• 2004

The NDVI(normalized difference vegetation index) dataset is unique or main tool to assess the global, multi seasonal, multi annual, and multi spectral changes over the World. These features are useful for environmental studies in particular, for the vegetation coverage monitoring of the country as Mongolia, where are large pastureland and pastoral animal husbandry, which dependent on natural conditions. Pasture vegetation cover is changing accordingly with both of global climate change and anthropogenic effect or human impacts. Using past 20 years (1982-2001) NDVI derived from NOAA satellite, its dynamical trend has been decreased in all natural zones differently. Also applied the method named "Two Years Differences" which could calculate the number of years with increased or decreased NDVI values at the same place. From May to September have occurred the 9 years maximum decreases of NDVI over Mongolia, but it obtained differently in spatial and temporal scale. In 24.4 ? 32.7% of all territory occurred one year decrease of NDVI and in 18% occurred more than 3 years frequent decrease of NDVI. According to the linear trend of NDVI and in 18% occurred more than 3 years frequent decrease of NDVI dynamics over 69% of whole territory of Mongolia NDVI values had been decreased due to both natural and human induced impacts to the pasture condition. In this paper also included some results of the integrated analyses of NOAA/NDVI and ground truth data over Monglia separately by natural zones.

• Advances in nano research
• /
• 제13권4호
• /
• pp.351-368
• /
• 2022

Recently, promising structural technologies like multi-function, ultra-load bearing capacity and tailored structures have been put up for discussions. Finite Element (FE) modelling is probably the best-known option capable of treating these superior properties and multi-domain behavior structures. However, advanced materials such as Functionally Graded Material (FGM) and nanocomposites suffer from problems resulting from variable material properties, reinforcement aggregation and mesh generation. Motivated by these factors, this research proposes a unified shape function for FGM, nanocomposites, graded nanocomposites, in addition to traditional isotropic and orthotropic structural materials. It depends not only on element length but also on the beam's material properties and geometric characteristics. The systematic mathematical theory and FE formulations are based on the Timoshenko beam theory for beam structure. Furthermore, the introduced element achieves C1 degree of continuity. The model is proved to be convergent and free-off shear locking. Moreover, numerical results for static and free vibration analysis support the model accuracy and capabilities by validation with different references. The proposed technique overcomes the issue of continuous properties modelling of these promising materials without discarding older ones. Therefore, introduced benchmark improvements on the FE old concept could be extended to help the development of new software features to confront the rapid progress of structural materials.

• Nuclear Engineering and Technology
• /
• 제50권1호
• /
• pp.54-67
• /
• 2018

There have been recent efforts to establish methods for high-fidelity and multi-physics simulation with coupled thermal-hydraulic (T/H) and neutronics codes for the entire core of a light water reactor under accident conditions. Considering the computing power necessary for a pin-by-pin analysis of the entire core, subchannel-scale T/H analysis is considered appropriate to achieve acceptable accuracy in an optimal computational time. In the present study, the applicability of in-house code CUPID of the Korea Atomic Energy Research Institute was extended to the subchannel-scale T/H analysis. CUPID is a component-scale T/H analysis code, which uses three-dimensional two-fluid models with various closure models and incorporates a highly parallelized numerical solver. In this study, key models required for a subchannel-scale T/H analysis were implemented in CUPID. Afterward, the code was validated against four subchannel experiments under unheated and heated single-phase incompressible flow conditions. Thereafter, a subchannel-scale T/H analysis of the entire core for an Advanced Power Reactor 1400 reactor core was carried out. For the high-fidelity simulation, detailed geometrical features and individual rod power distributions were considered in this demonstration. In this study, CUPID shows its capability of reproducing key phenomena in a subchannel and dealing with the subchannel-scale whole core T/H analysis.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권5호
• /
• pp.1431-1445
• /
• 2022

We construct a deep neural network model named ECGResNet. This model can diagnosis diseases based on 12-lead ECG data of eight common cardiovascular diseases with a high accuracy. We chose the 16 Blocks of ResNet50 as the main body of the model and added the Squeeze-and-Excitation module to learn the data information between channels adaptively. We modified the first convolutional layer of ResNet50 which has a convolutional kernel of 7 to a superposition of convolutional kernels of 8 and 16 as our feature extraction method. This way allows the model to focus on the overall trend of the ECG signal while also noticing subtle changes. The model further improves the accuracy of cardiovascular and cerebrovascular disease classification by using a fully connected layer that integrates factors such as gender and age. The ECGResNet model adds Dropout layers to both the residual block and SE module of ResNet50, further avoiding the phenomenon of model overfitting. The model was eventually trained using a five-fold cross-validation and Flooding training method, with an accuracy of 95% on the test set and an F1-score of 0.841.We design a new deep neural network, innovate a multi-scale feature extraction method, and apply the SE module to extract features of ECG data.

• Smart Structures and Systems
• /
• 제26권3호
• /
• pp.373-390
• /
• 2020

Hybrid simulation (HS) is a versatile tool for structural performance evaluation under dynamic loads. Although real structural responses are often multiple-directional owing to an eccentric mass/stiffness of the structure and/or excitations not along structural major axes, few HS in this field takes into account structural responses in multiple directions. Multi-directional loading is more challenging than uni-directional loading as there is a nonlinear transformation between actuator and specimen coordinate systems, increasing the difficulty of suppressing loading error. Moreover, redundant actuators may exist in multi-directional hybrid simulations of large-scale structures, which requires the loading strategy to contain ineffective loading of multiple actuators. To address these issues, lately a new strategy was conceived for accurate reproduction of desired displacements in bi-directional hybrid simulations (BHS), which is characterized in two features, i.e., iterative displacement command updating based on the Jacobian matrix considering nonlinear geometric relationships, and force-based control for compensating ineffective forces of redundant actuators. This paper performs performance validation and application of this new mixed loading strategy. In particular, virtual BHS considering linear and nonlinear specimen models, and the diversity of actuator properties were carried out. A validation test was implemented with a steel frame specimen. A real application of this strategy to BHS on a full-scale 2-story frame specimen was performed. Studies showed that this strategy exhibited excellent tracking performance for the measured displacements of the control point and remarkable compensation for ineffective forces of the redundant actuator. This strategy was demonstrated to be capable of accurately and effectively reproducing the desired displacements in large-scale BHS.

• 대기
• /
• 제22권3호
• /
• pp.331-344
• /
• 2012

The changes in the Hadley circulation during the second half of the 20th century were examined using observations and the 20C3M (Twentieth Century Climate in Coupled Models) simulations by the 21 IPCC AR4 models. Multi-model ensemble (MME) mean shows that the mean features of the Hadley circulation, such as the intensity, magnitude, and the seasonal variations, are very realistically reproduced, compared to the ERA40 reanalysis. But the long-term trends of the Hadley circulation in 20C3M MME are quite different to those of observations. The observed intensity of the Hadley cell is persistently enhanced, particularly during boreal winter. In comparison, the meridional overturning circulations reproduced in the MME mean remains invariant in time, and even weakened in boreal summer. This discrepancy between the ERA40 and 20C3M MME is consistently shown in the overall structure of the Hadley circulations, such as mass streamfunction, the velocity potential, the vertical shear of meridional wind, and the vertical velocity in the tropical region. This results indicate that the current climate models are skill-less to capture the long-term trend of Hadley circulation yet, and should be improved in simulation of the large-scale features to enhance the confidence level of future climate change projection.

• 융합신호처리학회논문지
• /
• 제11권2호
• /
• pp.143-150
• /
• 2010

무마커 증강현실 시스템은 실내나 옥외 환경에서 자연 물체를 인식하고 매칭하는 기능이 필수적이다. 본 논문에서는 비주얼 서술자와 코드북을 사용하여 특징을 추출하고 자연 물체를 인식하는 기법을 제안한다. 증강현실 응용은 동작 속도와 실시간 성능에 민감하기 때문에, 본 연구에서는 멀티 클래스의 자연 물체 인식에 초점을 두었으며 분류와 특징 추출 시간을 줄이는 것을 포함한다. 훈련과 테스트 과정에서 자연 물체로부터 특징을 추출하기 위해 SIFT와 SURF을 각각 사용하고 그들의 성능을 비교한다. 또한, 클러스터링 알고리즘을 이용하여 다차원의 특징 벡터들로부터 비주얼 코드북을 생성하고 나이브 베이즈 분류기를 이용해 물체를 인식한다.