• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권1호
• /
• pp.249-268
• /
• 2014

Relevance feedback is an effective tool to bridge the gap between superficial image contents and medically-relevant sense in content-based medical image retrieval. In this paper, we propose an interactive medical image search framework based on pairwise constraint propagation. The basic idea is to obtain pairwise constraints from user feedback and propagate them to the entire image set to reconstruct the similarity matrix, and then rank medical images on this new manifold. In contrast to most of the algorithms that only concern manifold structure, the proposed method integrates pairwise constraint information in a feedback procedure and resolves the small sample size and the asymmetrical training typically in relevance feedback. We also introduce a long-term feedback strategy for our retrieval tasks. Experiments on two medical image datasets indicate the proposed approach can significantly improve the performance of medical image retrieval. The experiments also indicate that the proposed approach outperforms previous relevance feedback models.

• 한국멀티미디어학회논문지
• /
• 제24권8호
• /
• pp.1122-1137
• /
• 2021

PM_2.5 concentration in Seoul could be predicted by deep neural network model. In this paper, the contribution of input factors to the model's prediction results is analyzed using the LRP(Layer-wise Relevance Propagation) technique. LRP analysis is performed by dividing the input data by time and PM concentration, respectively. As a result of the analysis by time, the contribution of the measurement factors is high in the forecast for the day, and those of the forecast factors are high in the forecast for the tomorrow and the day after tomorrow. In the case of the PM concentration analysis, the contribution of the weather factors is high in the low-concentration pattern, and that of the air quality factors is high in the high-concentration pattern. In addition, the date and the temperature factors contribute significantly regardless of time and concentration.

• 한국멀티미디어학회논문지
• /
• 제24권10호
• /
• pp.1414-1424
• /
• 2021

In this paper, the importance of input factors of a DNN (Deep Neural Network) PM_2.5 forecasting model using LRP(Layer-wise Relevance Propagation) is analyzed, and forecasting performance is improved. Input factor importance analysis is performed by dividing the learning data into time and PM_2.5 concentration. As a result, in the low concentration patterns, the importance of weather factors such as temperature, atmospheric pressure, and solar radiation is high, and in the high concentration patterns, the importance of air quality factors such as PM_2.5, CO, and NO₂ is high. As a result of analysis by time, the importance of the measurement factors is high in the case of the forecast for the day, and the importance of the forecast factors increases in the forecast for tomorrow and the day after tomorrow. In addition, date, temperature, humidity, and atmospheric pressure all show high importance regardless of time and concentration. Based on the importance of these factors, the LRP_DNN prediction model is developed. As a result, the ACC(accuracy) and POD(probability of detection) are improved by up to 5%, and the FAR(false alarm rate) is improved by up to 9% compared to the previous DNN model.

• 인터넷정보학회논문지
• /
• 제7권3호
• /
• pp.155-167
• /
• 2006

인터넷을 사용하는 사람들에게 그들의 관심사와 부합하는 웹 페이지를 제공하는 것은 매우 중요하다. 이러한 관점에서 본 논문은 각 웹 페이지의 주제와 연관된 정도를 계산하여 웹 페이지 군(cluster)을 형성하며, 단어빈도/문서빈도 엔트로피(entropy) 및 컴파일된 규칙을 이용하여 수집된 웹 페이지를 정제하는 주제 기반 웹 수집기를 제안한다. 실험을 통하여 주제 기반 웹 수집기에 대한 분류의 정확성, 수집의 효율성 및 수집의 일관성을 평가하였다. 첫째, C4.5, 역전패(back propagation) 및 CN2 기계학습 알고리즘으로 컴파일한 규칙을 이용하여 실험한 웹 수집기의 분류 성능은 CN2를 사용한 분류 성능이 가장 우수 하였으며, 둘째, 수집의 효율성을 측정하여 각 범주별로 최적의 주제 연관 정도에 대한 임계값을 도출할 수 있었다. 마지막으로, 제안한 수집기의 수집정도에 대한 일관성을 평가하기 위하여 서로 다른 시작 URL을 사용하여 수집된 웹 페이지들의 중첩정도를 측정하였다. 실험 결과에서 제안한 주제 기반 웹 수집기가 시작 URL에 큰 영향을 받지 않고 상당히 일관적인 수집을 수행함을 알 수 있었다.

• 한국수자원학회논문집
• /
• 제56권11호
• /
• pp.801-813
• /
• 2023

최근 인공신경망(Artificial Neural Network, ANN)의 연구가 활발하게 진행되면서 ANN을 이용하여 하천의 수질을 예측하는 연구가 진행되고 있다. 그러나 ANN은 Black-box의 형태이기 때문에 ANN 내부의 연산과정을 분석하는데 어려움이 있다. ANN의 연산과정을 분석하기 위해 설명가능한 인공지능(eXplainable Artificial Intelligence, XAI) 기술이 사용되고 있으나, 수자원 분야에서 XAI 기술을 활용한 연구는 미비한 실정이다. 본 연구는 XAI 기술 중 Layer-wise Relevance Propagation (LRP)을 사용하여 낙동강의 다산 수질관측소의 수온, 용존산소량, 수소이온농도 및 엽록소-a를 예측하기 위한 Multi Layer Perceptron (MLP)을 분석하였다. LRP를 기반으로 수질을 학습한 MLP를 분석하여 수질을 예측하기 위한 최적의 입력자료를 선정하고, 최적의 입력자료를 이용하여 학습한 MLP의 예측결과에 대한 분석을 실시하였다. LRP를 이용하여 최적의 입력자료를 선정한 결과를 보면, 수온, 용존산소량, 수소이온농도 및 엽록소-a 모두 주변지역의 일 강수량을 제외한 입력자료를 학습한 MLP의 예측정확도가 가장 높았다. MLP의 용존산소량 예측결과에 대한 분석결과를 보면, 최고점에서 수소이온농도 및 용존산소량의 영향이 크고 최저점에서는 수온의 영향이 큰 것으로 분석되었다.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제16권1호
• /
• pp.100-104
• /
• 2010

수많은 웹 문서로부터 웹 서퍼가 원하는 정보를 찾기 위해 다양한 검색 엔진들이 개발되어왔다. 검색 엔진에서 가장 중요한 기능 중 하나는 사용자 질의에 대해서 웹 문서를 평가하고 랭킹을 부여하는 것이다. PageRank등의 기존 하이퍼링크 정보를 이용한 웹 랭킹 알고리즘은 토픽 드리프트 현상을 발생시킨다. 이러한 문제를 해결하기 위하여 연관성 파급 모델이 제안되었지만, 기존의 연관성 파급 모델을 기반으로 하는 랭킹 알고리즘은 성능상의 이유로 실제 웹 검색 엔진에서 사용하기 어렵다. 본 논문에서는 이러한 토픽 드리프트 현상을 완화하면서 좋은 성능을 제공하는 새로운 랭킹 알고리즘을 제안한다. 다양한 실험을 통하여 기존 알고리즘들과 비교한 제안하는 알고리즘의 우수성을 검증한다.

• Structural Engineering and Mechanics
• /
• 제21권4호
• /
• pp.437-450
• /
• 2005

In this study, a complete analysis of soil-structure interaction problems is presented which includes a modelling of the near surrounding of the building (near-field) and a special description of the wave propagation process in larger distances (far-field). In order to reduce the computational effort which can be very high for time domain analysis of wave propagation problems, a special approach based on similarity transformation of the infinite domain on the near-field/far-field interface is applied for the wave radiation of the far-field. The near-field is discretised with standard Finite Elements, which also allows to introduce non-linear material behaviour. In this paper, a new approach to calculate the involved convolution integrals is presented. This approximation in time leads to a dramatically reduced computational effort for long simulation times, while the accuracy of the method is not affected. Finally, some benchmark examples are presented, which are compared to a coupled Finite Element/Boundary Element approach. The results are in excellent agreement with those of the coupled Finite Element/Boundary Element procedure, while the accuracy is not reduced. Furthermore, the presented approach is easy to incorporate in any Finite Element code, so the practical relevance is high.

• Geomechanics and Engineering
• /
• 제15권1호
• /
• pp.645-657
• /
• 2018

The paper aims to analyze the behaviour of torsional type surface waves propagating through fluid saturated inhomogeneous porous media clamped between two inhomogeneous anisotropic media. We considered three types of inhomogeneities in upper anisotropic layer which varies exponentially, quadratically and hyperbolically with depth. The anisotropic half space inhomogeneity varies linearly with depth and intermediate layer is taken as inhomogeneous fluid saturated porous media with sinusoidal variation. Following Biot, the dispersion equation has been derived in a closed form which contains Whittaker's function and its derivative, for approximate result that have been expanded asymptotically up to second term. Possible particular cases have been established which are in perfect agreement with standard results and observe that when one of the upper layer vanishes and other layer is homogeneous isotropic over a homogeneous half space, the velocity of torsional type surface waves coincides with that of classical Love type wave. Comparative study has been made to identify the effects of various dimensionless parameters viz. inhomogeneity parameters, anisotropy parameters, porosity parameter, and initial stress parameters on the torsional wave propagation by means of graphs using MATLAB. The study has its own relevance in connection with the propagation of seismic waves in the earth where fluid saturated poroelastic layer is present.

• Journal of Electrical Engineering and Technology
• /
• 제13권2호
• /
• pp.733-741
• /
• 2018

Capacity estimation is indispensable to ensure the safety and reliability of lithium-ion batteries in electric vehicles (EVs). Therefore it's quite necessary to develop an effective on-board capacity estimation technique. Based on experiment, it's found constant current charge time (CCCT) and the capacity have a strong linear correlation when the capacity is more than 80% of its rated value, during which the battery is considered healthy. Thus this paper employs CCCT as the health indicator for on-board capacity estimation by means of relevance vector machine (RVM). As the ambient temperature (AT) dramatically influences the capacity fading, it is added to RVM input to improve the estimation accuracy. The estimations are compared with that via back-propagation neural network (BPNN). The experiments demonstrate that CCCT with AT is highly qualified for on-board capacity estimation of lithium-ion batteries via RVM as the results are more precise and reliable than that calculated by BPNN.

• Computers and Concrete
• /
• 제25권4호
• /
• pp.343-354
• /
• 2020

In this paper, vibration characteristics of chiral double-walled carbon nanotubes entrenched on Kelvin's model. The Eringen's nonlocal elastic equations are being combined with Kelvin's theory to observe small scale response. A nonlocal model has been formulated to explore the frequency spectrum of chiral double-walled CNTs along with diversity of indices and nonlocal parameter. Wave propagation is proposed technique to establish field equations of model subjected to four distinct end supports. The significance of scale effect in relevance of length-to-diameter and thickness- to- radius ratios are discussed and displayed in detail.